)]}' { "log": [ { "commit": "2f9f3876dd51d6a3031220d578a18c98270c74a8", "tree": "2174a0aacd26955d17d7ac462c00769001226469", "parents": [ "5d40c67126d51ab1d78f44f51ec23cb75e9887fc" ], "author": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Wed Sep 29 19:48:08 2021 +0200" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Mon Oct 04 15:04:34 2021 +0000" }, "message": "m/n/b/unifiedefi: add bazel rule for EFI unified kernel images\n\nThis adds a new Bazel rule for generating EFI unified kernel images.\nIt also wraps the current Metropolis kernel in one for future use\nand as a cheap test.\n\nChange-Id: I74d7a25cc3c12d5f240eb218fbbd8ba09db1333a\nReviewed-on: https://review.monogon.dev/c/monogon/+/336\nReviewed-by: Sergiusz Bazanski \u003cserge@monogon.tech\u003e\n" }, { "commit": "bc7614ee942aee814865a5967c642802040476ed", "tree": "12573dc1932f0477dffec7a21cf635f8b3ea140a", "parents": [ "356b896eb4c3db9608d637c775845a09fc20fd07" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Thu Sep 09 13:07:09 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Wed Sep 29 13:04:33 2021 +0000" }, "message": "m/n/c/curator: fix watch fail on context timeout\n\nFixes https://github.com/monogon-dev/monogon/issues/75.\n\nChange-Id: Iefb772fa55499271e85fec500f50e6c77e49d05a\nReviewed-on: https://review.monogon.dev/c/monogon/+/326\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "356b896eb4c3db9608d637c775845a09fc20fd07", "tree": "af30addfbc8caba5275febf71847196f13aba8a5", "parents": [ "116c4a69dc90827d82023c362cbc26a17e188787" ], "author": { "name": "Mateusz Zalega", "email": "mateusz@monogon.tech", "time": "Tue Aug 10 17:27:15 2021 +0200" }, "committer": { "name": "Mateusz Zalega", "email": "mateusz@monogon.tech", "time": "Tue Sep 28 12:14:43 2021 +0000" }, "message": "m/n/b/mkverity: refactor into VerityEncoder\n\nThe implementation was refactored into a stream-oriented VerityEncoder and exposed for use outside the mkverity tool. In addition, end-to-end tests were provided.\n\nChange-Id: I2d009ca8030d6a86e9d6dbe6d6ae60a3b84d2d74\nReviewed-on: https://review.monogon.dev/c/monogon/+/314\nReviewed-by: Sergiusz Bazanski \u003cserge@monogon.tech\u003e\n" }, { "commit": "c72aa433ff68b76317233ba8805d55845e9efd1a", "tree": "ecdcb336a19c6cca8bbd426870b2ca023ec55207", "parents": [ "080f7ff710e359f2dab0d8ace98f0aa8e443d98d" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Wed Sep 22 09:35:16 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Mon Sep 27 18:54:36 2021 +0000" }, "message": "m/n/k/containerd: less verbose runsc logging\n\nThis disables debug logging on non -c dbg builds of Metropolis, making\nthe console output less verbose.\n\nThis should maybe just be \u0027fixed\u0027 by making logtree -\u003e console smarter,\nie. implement log output filters, but this will do for now - and I think\nin general we don\u0027t want to have super verbose per-syscall logging\nenabled unless necessary.\n\nThe implementation picks between two runsc .tomls to build the node\nimage with, which works for now but we might have to move to a\nfull-blown config generator (build-time or runtime) when we also start\nhaving differences between containerd.toml.\n\nChange-Id: I5539ca6f4763a769e879f7d50637aa36f3ffd29b\nReviewed-on: https://review.monogon.dev/c/monogon/+/327\nReviewed-by: Mateusz Zalega \u003cmateusz@monogon.tech\u003e\n" }, { "commit": "080f7ff710e359f2dab0d8ace98f0aa8e443d98d", "tree": "f515ece2d0503d897d2c09e18d9fcb5e2f57cee3", "parents": [ "44d6b832490adc28d787f392db1c9e40c9ff3438" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Thu Sep 09 13:01:00 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Tue Sep 21 08:41:43 2021 +0000" }, "message": "m/n/core/curator: add thin etcd storage abstraction\n\nThis implements etcdPrefix, a more formalized way to represent objects\nstored within etcd under some unique ID key.\n\nThis ensures any time objects are retrieved by key they are not\naccidentally traversing /-delimited \u0027path\u0027 elements, and implements the\nmildly complex range start/end computation operation for when all\nobjects from within a prefix must retrieved.\n\nChange-Id: Ib095f466faaf453b5f61a35642df6b0c1076ae05\nReviewed-on: https://review.monogon.dev/c/monogon/+/322\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "44d6b832490adc28d787f392db1c9e40c9ff3438", "tree": "f04b490de46c451931df9e078969eb550632e937", "parents": [ "e306d780504ae3ddfad3eb852c7adc5ec9757d89" ], "author": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Mon Sep 06 22:02:04 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Wed Sep 08 12:21:08 2021 +0000" }, "message": "Use a temporary ntp.monogon.dev CNAME instead of pool.ntp.org\n\nWe\u0027re currently trying to get a vendor zone assigned\n(monogon-dev/monogon#72). Meanwhile, use a CNAME pointing\nto pool.ntp.org to avoid a freak accident where someone uses\na WIP version of Metropolis to deploy a million plastic routers.\n\nChange-Id: Ib39006c65a23d2df3a1230c28b0b7245b9e3e3c4\nReviewed-on: https://review.monogon.dev/c/monogon/+/320\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "e306d780504ae3ddfad3eb852c7adc5ec9757d89", "tree": "3e8e2c7a21430777db525c9ed4717a2cab1c114a", "parents": [ "d7d6e0284de38cbeeb185ca17c0853b4b2c10ee9" ], "author": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Wed Sep 01 13:01:06 2021 +0200" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Mon Sep 06 09:05:40 2021 +0000" }, "message": "m/n/time: add time service\n\nThis adds a bare-minimum time service based on chrony/NTP for keeping\nthe system clock and RTC on Metropolis nodes accurate.\n\nIt also introduces a UID/GID registry in the Metropolis node code\nas this is the first unprivileged service to run on the node itself.\n\nIt does not yet use a secure time source, this is tracked as #73.\n\nChange-Id: I873971e6d3825709bc8c696e227bece4cfbda93a\nReviewed-on: https://review.monogon.dev/c/monogon/+/319\nReviewed-by: Sergiusz Bazanski \u003cserge@monogon.tech\u003e\n" }, { "commit": "d7d6e0284de38cbeeb185ca17c0853b4b2c10ee9", "tree": "37e0b443caf904f0b78d423ba6580c1416f5bc11", "parents": [ "9ffa1f9577003ab70a6b483475874f3552d1ccc3" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Wed Sep 01 15:03:06 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Fri Sep 03 11:15:40 2021 +0000" }, "message": "m/n/core/rpc: create library for common gRPC functions\n\nThis is the beginning of consolidating all gRPC-related code into a\nsingle package.\n\nWe also run the Curator service publicly and place it behind a new\nauthorization permission bit. This is in preparation for Curator\nfollowers needing access to this Service.\n\nSome of the service split and authorization options are likely to be\nchanged in the future (I\u0027m considering renaming Curator to something\nelse, or at least clearly stating that it\u0027s a node-to-node service).\n\nChange-Id: I0a4a57da15b35688aefe7bf669ba6342d46aa3f5\nReviewed-on: https://review.monogon.dev/c/monogon/+/316\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "9ffa1f9577003ab70a6b483475874f3552d1ccc3", "tree": "a688d02424e8601ed830d12021b5867688d31438", "parents": [ "6bd415920b84bd695038caeb386f1e97184f0c51" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Wed Sep 01 15:42:23 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Thu Sep 02 10:38:15 2021 +0000" }, "message": "m/n/core/curator: authenticated RPC\n\nThis adds authentication middleware (server interceptors) for gRPC\nservices running on the public curator listener.\n\nMost of this code is testing harnesses to start up just the curator\nlistener with enough of a PKI infrastructure copy from a real Metropolis\ncluster to be able to start running tests against GetRegisterTicket.\n\nChange-Id: I429ff29e3c1233d74e8da619ddb543d56bc051b9\nReviewed-on: https://review.monogon.dev/c/monogon/+/311\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "6bd415920b84bd695038caeb386f1e97184f0c51", "tree": "60a125e0a299663c392218ed8b80cf33ea31aabd", "parents": [ "68dcee136984e2e16b7682e0c0758c1df831a84c" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Mon Aug 23 13:18:37 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Wed Sep 01 12:56:15 2021 +0000" }, "message": "m/node: add Management service, implement GetRegisterTicket RPC\n\nThis follows the Cluster Lifecycle design document.\n\nDO NOT MERGE: this needs a stacked CL on top which implements\nauthentication for the Management service.\n\nChange-Id: I19422a63b9dbf2fc0c7f4cbe204851af35b4dbdf\nReviewed-on: https://review.monogon.dev/c/monogon/+/307\nReviewed-by: Mateusz Zalega \u003cmateusz@monogon.tech\u003e\n" }, { "commit": "031243f5a276726080a92410f7d3503e5870ed49", "tree": "ab582e63dccf71c27e916d23ea24d5f250774d41", "parents": [ "cbeb8a01de2ac264f41b403b6fdc33dca7b5e568" ], "author": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Tue Aug 24 12:14:27 2021 +0200" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Thu Aug 26 16:02:23 2021 +0000" }, "message": "m/p/devicemapper: fix GC closing control fd\n\nThe devicemapper package stored a reference to its control file\ndescriptor as a uintptr after opening it thorugh os.Open(). This is a\nproblem as os.newFile (internally called by os.Open) sets a finalizer\non the os.File which closes the fd as soon as the object is GCed.\nBecause no such reference was kept by the devicemapper package, the GC\ncould end up closing the fd.\n\nTo fix this, the package now keeps the original os.File around and\njust grabs an Fd as necessary. While we\u0027re at it, let\u0027s make the\ncontrol file descriptor implementation threadsafe.\n\nChange-Id: I6f7e0a398f28c1141627904ccbd2d99dd248bc78\nReviewed-on: https://review.monogon.dev/c/monogon/+/310\nReviewed-by: Sergiusz Bazanski \u003cserge@monogon.tech\u003e\nVouch-Run-CI: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "c1bf6aa7ac83513659d56756009d572deffa7177", "tree": "fafe3258a74a97171a9d11d917a19c2e7387db73", "parents": [ "1f9a03b3f952320824b1ae49e56da3cb814cd5b0" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Mon Aug 23 13:05:24 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Tue Aug 24 17:26:35 2021 +0000" }, "message": "m/n/core/curator: fix listener stuckness on restarts\n\nThis makes both gRPC listener runnables (local and public) manage their\nown listening sockets, allowing them to restart independently of\neachother, and making sure that any listening sockets are cleaned up.\n\nWe also fix the existing curator test (which does not exercise the\nlisteners, just leadership election) to place the curators and their\nlocal sockets in /tmp instead of the default bazel tempdir (as a path\nbased on that is longer than the maximum domain socket path). This makes\nthese tests slightly less noisy (as they kept crashing while not being\nable to listen to the local socket).\n\nThis should\u0027ve been caught by a curator listener test, if we had one\n(other than the e2e test). I\u0027m growing keen on spending some time\nwriting enough of a harness to actually do that. Maybe once we have a\nfollower implementation ready…\n\nChange-Id: I0267292781b6ee8aff1d0557d420bbaa3c3d79f6\nReviewed-on: https://review.monogon.dev/c/monogon/+/304\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "1f9a03b3f952320824b1ae49e56da3cb814cd5b0", "tree": "315ea3ca5711b2dca9173dcf825c18e031affa84", "parents": [ "b9044c888097757c36933062f27b5f5ee103ee5f" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Tue Aug 17 13:40:53 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Tue Aug 24 17:26:20 2021 +0000" }, "message": "m/test/e2e: retrieve owner credentials in e2e test\n\nThis exercises AAA.Escrow for the initial cluster owner within our large\ne2e test suite. The certificate retrieved this way is not yet used, but\nis verified to be emitted for the correct public key.\n\nChange-Id: Id33178cd223e3180d6f834c6fac94d6d657d5349\nReviewed-on: https://review.monogon.dev/c/monogon/+/290\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "b9044c888097757c36933062f27b5f5ee103ee5f", "tree": "b07722231a9cf0fd3c0b81486bd637e11cbd7b6b", "parents": [ "3bb23219009a98643a562b1e59e3a4080f422c51" ], "author": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Tue Aug 24 11:59:47 2021 +0200" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Tue Aug 24 16:12:52 2021 +0000" }, "message": "m/p/devicemapper: make parameter encoding part of package\n\nThe DM kernel interface gets a single parameter string for each DM\ntarget in a table but internally the kernel immediately decodes it into\nan argv-style list of string arguments. Because everything needs to do\nit and it can be quite hard to get right, let\u0027s make it part of the\ndevicemapper package. Properly encoding this also means you get\nactionable errors when you pass invalid data instead of weird kernel\nerrors or misbehavior.\n\nChange-Id: I8060871a7459183c0395e5e4e8aac517544b2e87\nReviewed-on: https://review.monogon.dev/c/monogon/+/309\nReviewed-by: Sergiusz Bazanski \u003cserge@monogon.tech\u003e\n" }, { "commit": "41d275a63864e67deacad5b2ec0b435b01984034", "tree": "b5acf06483c0c33f0e988c82cbb25db4ceea777f", "parents": [ "5b2ae5500a90dc48b9713095e5f1580b9c9646d9" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Tue Aug 17 13:09:43 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Tue Aug 24 13:18:12 2021 +0000" }, "message": "m/n/c/curator: implement AAA.Escrow for initial owner pubkey\n\nThis finally implements AAA.Escrow in Metropolis.\n\nWe\u0027re not yet implementing multi-user support, so this currently only\nimplements retrieving an Owner certificate using the owner public key\nspecified in NodeParameters.cluster_bootstrap.\n\nChange-Id: I64a7ba025a8069d82b3c804ca3e2a706de2b0fbf\nReviewed-on: https://review.monogon.dev/c/monogon/+/289\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "5b2ae5500a90dc48b9713095e5f1580b9c9646d9", "tree": "1f6efbed2aa20716c18772bb30dbafacd6f07db3", "parents": [ "03758714f4b7be2a712831beecfdfcbf151b4c66" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Tue Aug 17 13:00:14 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Mon Aug 23 12:26:16 2021 +0000" }, "message": "m/n/c/curator: listen on public gRPC\n\nThis enables listening on CuratorPort (which was called\nNodeServicePort) using TLS node certificates. No service is yet running\non the new gRPC listener.\n\nChange-Id: I436ac1ae9cbdb257419ad114262fc2a7516396b1\nReviewed-on: https://review.monogon.dev/c/monogon/+/288\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "03758714f4b7be2a712831beecfdfcbf151b4c66", "tree": "1a7e2f3096a130897f53c28d4f9b72ea871264d0", "parents": [ "8ff4b7c6f20c9dda91c0eefc524e9bb6c3bff52d" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Tue Aug 17 12:52:11 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Fri Aug 20 09:03:18 2021 +0000" }, "message": "m/n/core: save owner public key in etcd\n\nThis is an early implementation of storing user credentials. It\ncurrently does not support more then the owner credentials.\n\nThese are not yet used anywhere, but will be in a follow-up CL.\n\nChange-Id: Ib876f7aaff44531dcae5a27875a960aaa9ec029f\nReviewed-on: https://review.monogon.dev/c/monogon/+/287\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "158e9a415a72bfacfdf9f46eb06b30486680299f", "tree": "ef5884479bc8da914cb69f7c111d06e32cbe2a1e", "parents": [ "a41caacc71418f7307d851fad95991cf80bdcb41" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Tue Aug 17 17:04:54 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Thu Aug 19 11:52:17 2021 +0000" }, "message": "m/n/core: reformat\n\nWe should get this into CI, as my IntelliJ keeps finding Go files that\naren\u0027t properly reformatted.\n\nChange-Id: I7ea736994bea13ccebc787548fadf3da984e334e\nReviewed-on: https://review.monogon.dev/c/monogon/+/291\nReviewed-by: Lorenz Brun \u003clorenz@monogon.tech\u003e\n" }, { "commit": "a41caacc71418f7307d851fad95991cf80bdcb41", "tree": "cbcf9af76f29ccb94b7c2b94d75f1e8eb39cfb3d", "parents": [ "5253884d51cb64c1d1afcb2d7b969f7c2b50b302" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Thu Aug 12 17:00:55 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Thu Aug 19 10:20:55 2021 +0000" }, "message": "m/pkg/pki: forbid External/Managed certificates without name\n\nThis ensures any stored certificates must have a name set - otherwise\nthey end up being created with an empty string as a name, and end up\ncolliding with eachother.\n\nChange-Id: I9e415b6ff89dbda179526920d58e33e638a28cec\nReviewed-on: https://review.monogon.dev/c/monogon/+/286\nReviewed-by: Mateusz Zalega \u003cmateusz@monogon.tech\u003e\n" }, { "commit": "5253884d51cb64c1d1afcb2d7b969f7c2b50b302", "tree": "10a6bf03472e9c14da2515ea7755d74bb3f660e6", "parents": [ "99f477412a2e701f89f7698be1dd432adcfff17c" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Wed Aug 11 16:22:41 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Thu Aug 19 10:20:55 2021 +0000" }, "message": "m/pkg/pki: refactor, allow for external certificates\n\nThe pki library supported managing certificates in two modes:\n\n - default, when name !\u003d \"\"\n - volatile/ephemeral, when name \u003d\u003d \"\"\n\nThe difference between the two being that default certificates were\nfully stored in etcd (key and x509 certificate), while volatile\ncertificates weren\u0027t stored at all. However, both kinds needed private\nkeys passed to the pki library.\n\nWe want to be able to emit certificates without having private keys for\nthat certificate, so we end up a third mode of operation: \u0027external\ncertificates\u0027. These are still stored in etcd, but without any\ncorresponding private key.\n\nIn the future we might actually get rid of ephemeral certificates by\nexpanding the logic of external certificates to provide a full audit log\nand revocation system, instead of matching by Certificate Name. But this\nwill do for now.\n\nWe also use this opportunity to write some simple tests for this\npackage.\n\nChange-Id: I193f4b147273b0a3981c38d749b43362d3c1b69a\nReviewed-on: https://review.monogon.dev/c/monogon/+/263\nReviewed-by: Mateusz Zalega \u003cmateusz@monogon.tech\u003e\n" }, { "commit": "99f477412a2e701f89f7698be1dd432adcfff17c", "tree": "43c088e34cfa9171a5587573f4d824f9d09e0a69", "parents": [ "9d6c4c78bca9da0db7e40e5de776528f3d83a7dd" ], "author": { "name": "Serge Bazanski", "email": "serge@monogon.tech", "time": "Wed Aug 04 20:21:42 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@monogon.tech", "time": "Thu Aug 19 10:20:55 2021 +0000" }, "message": "curator: provisions for implementing multiple gRPC services\n\nWe want to run some other gRPC services on the Curator leader/follower\nimplementations other than just the Curator gRPC service.\n\nThis decouples the local types from implementing a particular gRPC\nservice (instead proxying through an interface) and splits out the\nimplementation of the Curator gRPC service from the main leader objects.\n\nThis should allow us to add an implementation of eg. a Management gRPC\nservice in a testable manner (the only thing we have to dependency\ninject is the leadership struct, and that\u0027s trivial to do with a simple\netcd test server).\n\nChange-Id: Ia0ea65e40a775bf49661d0b99c0185aa83547ed0\nReviewed-on: https://review.monogon.dev/c/monogon/+/260\nReviewed-by: Mateusz Zalega \u003cmateusz@monogon.tech\u003e\n" }, { "commit": "9d6c4c78bca9da0db7e40e5de776528f3d83a7dd", "tree": "71cba058637a866b7c87e3f296dd4995d0171cfe", "parents": [ "257acab41f5a35575ca0f2dbc9568b1bd75d2570" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jul 20 21:16:27 2021 +0200" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Wed Aug 11 15:38:58 2021 +0000" }, "message": "Support injecting container images for development\n\nThis adds the LoadImage RPC and an accompanying subcommand\nto the debug API which allows loading images into\nan existing Metropolis node for\ndevelopment or testing.\n\nChange-Id: I51d802630ae4c95fb874e01bfb6510ab69c322e1\nReviewed-on: https://review.monogon.dev/c/monogon/+/219\nReviewed-by: Sergiusz Bazanski \u003cserge@monogon.tech\u003e\n" }, { "commit": "257acab41f5a35575ca0f2dbc9568b1bd75d2570", "tree": "fdc41d8de424f74525b7a92024c12f00ed8928fa", "parents": [ "1445396219351e711f82d4cebad6e84a46553bda" ], "author": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Tue Aug 10 12:36:17 2021 +0200" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@monogon.tech", "time": "Wed Aug 11 11:28:06 2021 +0000" }, "message": "m/p/devicemapper: Support creating read-only devices\n\nI originally thought this is not going to be needed as R/W control can be done through devicemapper itself, but verity requires a read-only table.\n\nWhile we\u0027re here let\u0027s also add some doc comments to the Target struct.\n\nExisting functionality is covered by existing tests, read-only functionality will be exercised by verity tests once they land.\n\nChange-Id: Ib76bcffb14b5fe40d8d77bd9731b591d0d8cf22f\nReviewed-on: https://review.monogon.dev/c/monogon/+/262\nReviewed-by: Sergiusz Bazanski \u003cserge@monogon.tech\u003e\n" }, { "commit": "1445396219351e711f82d4cebad6e84a46553bda", "tree": "798347e10dba644f861fd9dbf55fe59cef9fb82f", "parents": [ "6767e052c761f2b19a4966f707c65d8bc08c3c3c" ], "author": { "name": "Mateusz Zalega", "email": "mateusz@monogon.tech", "time": "Fri Jul 23 16:58:02 2021 +0200" }, "committer": { "name": "Mateusz Zalega", "email": "mateusz@monogon.tech", "time": "Tue Aug 10 14:39:14 2021 +0000" }, "message": "m/n/b/mkverity: implement a dm-verity hash image generator\n\nBackground: https://github.com/monogon-dev/monogon/issues/57\n\nThe piece of code included implements a subset of veritysetup\nfunctionality (see: dm-verity). It was written in an attempt to\nminimize projected higher maintenance cost of packaging cryptsetup\nfor metropolis in the long term.\n\nThe implementation was verified with the original veritysetup tool:\n\u003e$ ./go-veritysetup format file1 file2\n\u003e33359c1f1bdd25e7afc2e98cd27c440e7af9ef2fb55462ce562a1b8254bf02e4\n\u003e$ veritysetup --debug --verbose verify file1 file2 33359c1f1bdd25e7afc2e98cd27c440e7af9ef2fb55462ce562a1b8254bf02e4\n\nKtest-based tests and buildsystem integration are still pending.\n\nCompatibility with the original cryptsetup tool might be dropped\neventually, if it\u0027s found beneficial to do so.\n\nChange-Id: I5a6e1b18b692b1701e405013f132f6f2711b2c96\nReviewed-on: https://review.monogon.dev/c/monogon/+/250\nReviewed-by: Sergiusz Bazanski \u003cserge@monogon.tech\u003e\n" }, { "commit": "439b95eb515c86ba8ce9917da258c0875f36f038", "tree": "b7590aaac27700868ca1638af4d2e84bfb1a975f", "parents": [ "f9edf520bd3fa049b8f1d5f09f0ee727864ca8c7" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Jun 30 23:16:13 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Tue Jul 20 14:43:28 2021 +0000" }, "message": "m/n/c/{cluster,curator}: move NodeCredentials to cluster\n\nThis keeps the NodeCredentials/NodeCertificate logic types near their\nconsumer, the cluster bootstrap code. It also rewrites these structures\nto be centered around the x509 data itself.\n\nThis is a followup to https://review.monogon.dev/c/monogon/+/186 .\nAttempting to introduce it into that change was too complex due to the\nsurrounding cluster/curator refactoring.\n\nWe also take this opportunity to write some simple tests for the\ncredential validation logic.\n\nChange-Id: Iead3cfdd5778274508d79799f4750f5fdf9385bc\nReviewed-on: https://review.monogon.dev/c/monogon/+/197\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "f9edf520bd3fa049b8f1d5f09f0ee727864ca8c7", "tree": "1fd016dcbeccbc7b8bbfc32960c43bd0dbaffe5f", "parents": [ "a959cbd12c29b62045f02b1d22cdf1e36c0261a4" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jun 17 15:57:13 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Tue Jul 20 14:42:45 2021 +0000" }, "message": "metropolis/node/core: use curator\n\nThis finally switches over the node startup code to use the full Cluster\nManager / Curator / Role Server chain to bring up the node.\n\nChange-Id: Iaf6173671aed107a67b4201d9d1ad8bb33baa90f\nReviewed-on: https://review.monogon.dev/c/monogon/+/189\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "a959cbd12c29b62045f02b1d22cdf1e36c0261a4", "tree": "7f44cdf24c3c02f7d9ba866a2d37275673a0dd11", "parents": [ "0d93777cf32dd0d0f6f8d75d8396f7140cff9d13" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jun 17 15:56:51 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Tue Jul 20 12:02:29 2021 +0000" }, "message": "metropolis/node/core/cluster: use curator\n\nThis refactors the cluster manager. It removes all etcd storage\nfunctionality (which now lives in the curator) and otherwise dusts\nthings off slightly (some file renames, some comments to reflect the now\nclarified and limited scope of the cluster manager).\n\nChange-Id: Ic62d8402c0618fb5e0e65966b0d732a2cab564e0\nReviewed-on: https://review.monogon.dev/c/monogon/+/188\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "0d93777cf32dd0d0f6f8d75d8396f7140cff9d13", "tree": "f6abee66c82ba141ad449575a6d35d06408c6ed4", "parents": [ "963c4090f74341d8efc61b49ba5934a18434371c" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jun 17 15:54:40 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Tue Jul 20 12:02:29 2021 +0000" }, "message": "m/n/core/roleserver: implement\n\nThis implements the Role Server, which is the new service responsible\nfor actually running Metropolis workloads like the Kubernetes services.\n\nThis decouples starting Kubernetes from node startup code, and handles\ncases like nodes changing roles at runtime.\n\nChange-Id: Ie5f7f2c30b05fe74ca003805532dba50bf3821d0\nReviewed-on: https://review.monogon.dev/c/monogon/+/187\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "963c4090f74341d8efc61b49ba5934a18434371c", "tree": "65e0e58e997e155290118e53aebfc228aa3e20be", "parents": [ "2098b98c7deaf9115742cf73071f888e0513cf2f" ], "author": { "name": "Mateusz Zalega", "email": "mateusz@monogon.tech", "time": "Mon Jul 05 18:50:08 2021 +0200" }, "committer": { "name": "Mateusz Zalega", "email": "mateusz@monogon.tech", "time": "Mon Jul 19 10:52:16 2021 +0000" }, "message": "m/n/c/consensus_test: move timeout handling to Bazel\n\nThe test regularly exceeds timeout on slower machines, and besides,\nthat\u0027s the proper way to do it.\n\nChange-Id: Ic4a8ac717bd7f6e70d4d4ac0b156f42ff5addef8\nReviewed-on: https://review.monogon.dev/c/monogon/+/215\nReviewed-by: Leopold Schabel \u003cleo@nexantic.com\u003e\nReviewed-by: Sergiusz Bazanski \u003cserge@nexantic.com\u003e\n" }, { "commit": "2098b98c7deaf9115742cf73071f888e0513cf2f", "tree": "6037aec601525299a09d8996f4ebe0c1e4a91674", "parents": [ "79fc1e9fd6ee8777f097ab251b828d82e33b7bad" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 15:13:46 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 15:34:42 2021 +0000" }, "message": "m/pkg/combinectx: reformat\n\nSeems like this slipped past the cracks on original review - we should\nadd CI for this.\n\nChange-Id: I35cc1d14710109d4d2d0a60b573400b65cb7d350\nReviewed-on: https://review.monogon.dev/c/monogon/+/212\nReviewed-by: Mateusz Zalega \u003cmateusz@monogon.tech\u003e\n" }, { "commit": "79fc1e9fd6ee8777f097ab251b828d82e33b7bad", "tree": "a1c50daa06f2ea66fb5fcf14e85331385facdb68", "parents": [ "50009e024b50eda2c69b884600d0850c73d62b6d" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Jul 06 16:25:22 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 13:49:55 2021 +0000" }, "message": "metropolis/*: move tests to supervisor harness\n\nThis somewhat simplifies test code (barely), but more importantly pipes\nlogs from runnables into test stdout, thereby making debugging much\neasier.\n\nChange-Id: I3e597bbac8497bea3477afd54f61b592a0d08355\nReviewed-on: https://review.monogon.dev/c/monogon/+/206\nReviewed-by: Leopold Schabel \u003cleo@nexantic.com\u003e\n" }, { "commit": "50009e024b50eda2c69b884600d0850c73d62b6d", "tree": "28ee724fd39c537e09a755fbd593dc15d31e288e", "parents": [ "ebe025936fc86f53e7316f894f54dd6ef9b0a9d7" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 14:35:27 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 13:48:54 2021 +0000" }, "message": "m/n/core/consensus: parse etcd server logs\n\nThis finally gives us easy to read etcd logs instead of raw JSON dumps\ninto stdout. Instead of simply parsing them as raw logs, we convert them\ninto leveled logs.\n\nChange-Id: I7cfe18b9c4e24d7742a01a77f5d9c6ddee647493\nReviewed-on: https://review.monogon.dev/c/monogon/+/209\nReviewed-by: Leopold Schabel \u003cleo@nexantic.com\u003e\n" }, { "commit": "ebe025936fc86f53e7316f894f54dd6ef9b0a9d7", "tree": "0dd0a48c297e69a8bcbe53ef65d3dba7f53961a3", "parents": [ "020b7c53a59f7f4e31976d7b3f08011dadb1c9c4" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 14:23:26 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 13:45:29 2021 +0000" }, "message": "m/pkg/logtree/unraw: implement\n\nThis is another part of the generic external leveled log ingestion\nmechanism. This parts takes care of ingesting external data either by\nexposing an io.Writer or a named pipe on the filesystem from which\nexternal logs are parsed and then inejcted into the logtree.\n\nChange-Id: Ie2263496ca4d50220abdd8e4d37a35730d127319\nReviewed-on: https://review.monogon.dev/c/monogon/+/208\nReviewed-by: Leopold Schabel \u003cleo@nexantic.com\u003e\n" }, { "commit": "020b7c53a59f7f4e31976d7b3f08011dadb1c9c4", "tree": "9c6b8ea68b0a4db7d4a8b90b636feff712998235", "parents": [ "f8a8e65685cb621dc7fb39043a6d01caee5dcaf0" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 14:22:28 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 13:45:21 2021 +0000" }, "message": "metropolis/pkg/logtree: allow logging external leveled payloads\n\nThis is in preparation for making the mechanism to ingest external\nlogging more generic (currently we have an ad-hoc solution for klog, but\nwe now also want to implement one for etcd).\n\nChange-Id: I6e6f656e5d83ad22d67a81fbeb87c8d369796e18\nReviewed-on: https://review.monogon.dev/c/monogon/+/207\nReviewed-by: Leopold Schabel \u003cleo@nexantic.com\u003e\n" }, { "commit": "f8a8e65685cb621dc7fb39043a6d01caee5dcaf0", "tree": "db6142898e003969628a3ec879f6af77780f8da4", "parents": [ "f0b4da54afc17f4b2b1c31ddb9433ee888aea699" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Jul 06 16:23:43 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 13:36:48 2021 +0000" }, "message": "m/pkg/{logtree,supervisor}: add test helpers\n\nThis adds two functions:\n\n logtree.PipeAllToStderr\n supervisor.NewHarness\n\nThese are designed to simplify tests that exercise code which expects to\nbe run as a supervisor runnable and/or have access to a logtree\ninstance.\n\nChange-Id: Ibce77aa4927515af7c273d07ced15215ff456ecc\nReviewed-on: https://review.monogon.dev/c/monogon/+/205\nReviewed-by: Leopold Schabel \u003cleo@nexantic.com\u003e\n" }, { "commit": "f0b4da54afc17f4b2b1c31ddb9433ee888aea699", "tree": "a4ef4b8154c4024714209710696310bfd7bc2764", "parents": [ "35e43d133a16750adfa1683473f5c2648a010b1a" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Mon Jun 21 20:05:59 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 13:31:45 2021 +0000" }, "message": "m/n/core/curator: implement leader and Node/PKI state\n\nThis implements actual Curator logic for nodes and PKI. These will\nreplace the cluster manager\u0027s equivalent logic.\n\nThere are two entry points to this logic:\n\n - the gRPC service\u0027s Watch method for accessing node status\n - bootstrap logic to create a node when the cluster manager bootstraps\n the cluster.\n\nTest plan: a followup CR will introduce tests for the Curator - more\ngranular than the full E2E suite. DO NOT MERGE UNTIL THEN, as this is\ncritical code.\n\nChange-Id: I8c40a821b846012b90cf9a5df27901d1b49f388c\nReviewed-on: https://review.monogon.dev/c/monogon/+/186\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "35e43d133a16750adfa1683473f5c2648a010b1a", "tree": "6aa1e8bcebd03a74b3950128436c5a37268d87c0", "parents": [ "3c885deeda9ab560ee29e94159782ce4323af80e" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Jul 06 13:12:14 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Wed Jul 07 13:24:21 2021 +0000" }, "message": "m/pkg/supervisor: move internal testhelpers\n\nThese are helper functions used for internal supervisor tests. This move\nis in preparation for writing the other kind of \u0027test helers\u0027: ones that\nare used by tests in other libraries when interacting with supervisor\ntypes.\n\nChange-Id: I64efe19b68c7c244ad426167565b0083a1b86fcf\nReviewed-on: https://review.monogon.dev/c/monogon/+/204\nReviewed-by: Leopold Schabel \u003cleo@nexantic.com\u003e\n" }, { "commit": "3c885deeda9ab560ee29e94159782ce4323af80e", "tree": "5d6a18c0b822accc5c279240c0d7e52ca071a361", "parents": [ "aad79488d417eefafd4102bda9bd10d6473cf6c7" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jun 17 17:21:00 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Mon Jul 05 18:19:36 2021 +0000" }, "message": "m/n/core/curator: implement gRPC listener\n\nThis implements the Curator listener and listener dispatcher, two\nrunnables responsible for maintaining an active Curator RPC\nimplementation (either leader or follower) and switching over\nappropriately as the election status changes.\n\nThis might be overengineered. The implementation switchover logic and\ncontext joining could possibly be ripped out and replaced by plain gRPC\nsocket close and re-open on switchover.\n\nTesting plan: implements unit tests for switching/dispatching.\n\nChange-Id: Ib62195b336d0754d99865d7a2a32ba2ffa3543ce\nReviewed-on: https://review.monogon.dev/c/monogon/+/185\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "aad79488d417eefafd4102bda9bd10d6473cf6c7", "tree": "a638f94700704699b4ae4a5ad6b5214d6adcb479", "parents": [ "cbf1fa97307024b1f0c60c88e8ebf968a42bf980" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri Jul 02 17:40:36 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Mon Jul 05 18:19:29 2021 +0000" }, "message": "metropolis/handbook: init\n\nThis is the beginning of the Metropolis Handbook, the main end-user\ndocumentation of Metropolis.\n\nIt is built using mdbook, and currently contains only the default\ncontent from `mdbook init`.\n\nFuture work: start writing the handbook, build in CI, publish in CI.\n\nChange-Id: I81753350215b2f7aabc17925eadfd20706e1fdb5\nReviewed-on: https://review.monogon.dev/c/monogon/+/202\nReviewed-by: Leopold Schabel \u003cleo@nexantic.com\u003e\n" }, { "commit": "76003f807b24a22476b14bc308939fc62e1ad6a2", "tree": "24fc636b5c47569aa8e64119bd4b1f81a794c9ab", "parents": [ "e30d7d08548fe233cacff5870971f08318ff9646" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jun 17 16:39:01 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Wed Jun 30 12:29:59 2021 +0000" }, "message": "m/n/core/curator: implement leader election\n\nThis implements the leader election functionality subset of the curator.\nIt does not yet implement any business logic, just the switchover\nbetween acting as a leader and a follower.\n\nTest plan: implements an integration test for the leader election with\nan in-memory etcd cluster.\n\nChange-Id: Id77ecc35a9f2b18e716fffd3caf2de193982d676\nReviewed-on: https://review.monogon.dev/c/monogon/+/184\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "e30d7d08548fe233cacff5870971f08318ff9646", "tree": "be4b12585cf856db8fc74746f5bfed69748195cd", "parents": [ "db77e82767ff620de89d75a7f3e9a6c37b1d0973" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Jun 23 18:50:19 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Thu Jun 24 09:08:20 2021 +0000" }, "message": "m/n/core/consensus/client: add context to ThinClient\n\nThis context ends up in the returned clientv3.Client structure and is\naccessible under .Ctx(). This is used by some library code, and all\netcd clients by default have a context, so that library code will fail\nif none is set.\n\nChange-Id: If2b9fb0720725d623c05ea4947fdf99fe6d5d1d0\nReviewed-on: https://review.monogon.dev/c/monogon/+/195\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "db77e82767ff620de89d75a7f3e9a6c37b1d0973", "tree": "f6e30c8aaae105c02324a16e640f5de97a3be2c7", "parents": [ "d98ad45e64b542f9945939d35cba9ebd352ff913" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jun 08 14:08:34 2021 +0200" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jun 22 12:06:18 2021 +0000" }, "message": "Protocol specification for VMs\n\nThese are the protobufs for the VM system communication\n\nChange-Id: I16915c5391a05b7c7a9c68a0017dc2365de7e367\nReviewed-on: https://review.monogon.dev/c/monogon/+/125\nReviewed-by: Sergiusz Bazanski \u003cserge@nexantic.com\u003e\n" }, { "commit": "d98ad45e64b542f9945939d35cba9ebd352ff913", "tree": "cd8fa82c4ccd2bf0616986615f3eb2d7c4798f30", "parents": [ "7f17d9b41f248f4b009f5d702622616f62d0a2fa" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jun 17 15:55:17 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Tue Jun 22 12:05:20 2021 +0000" }, "message": "metropolis/n/kubernetes/pki: remove verbose \u0027ensure\u0027 cert logging\n\nChange-Id: Ie12f12a2966282d364730a1c7a148fae78ab236d\nReviewed-on: https://review.monogon.dev/c/monogon/+/190\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "7f17d9b41f248f4b009f5d702622616f62d0a2fa", "tree": "fdb69964144f9b654a746d15e86d34f0bf1b108b", "parents": [ "63fabd954a5303f6bba41557da55b1940198058f" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jun 17 16:11:40 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Tue Jun 22 12:05:16 2021 +0000" }, "message": "m/n/core/localstorage: add curator ephemeral directory\n\nThis is where the curator will listen for local gRPC connections from\nthe rest of the node code.\n\nChange-Id: Id0ea2e4e05d04bbea4b3b88dd51122d394317578\nReviewed-on: https://review.monogon.dev/c/monogon/+/183\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "63fabd954a5303f6bba41557da55b1940198058f", "tree": "547628727ea80a09eced09a402945d81263f6778", "parents": [ "30653eeb3adcbebfecb4216569dc247b51cfb464" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jun 17 15:47:22 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Mon Jun 21 09:12:58 2021 +0000" }, "message": "m/n/c/curator/proto: add API proto\n\nThis is the initial API of the curator service we\u0027re about to implement.\nThe TODO comments already reflect the state of the implementation as is\nnow, and as will be stacked on top of this change.\n\nTesting Plan: proto change only, no logic yet.\n\nChange-Id: I2c11b0d5f2112b7872ea348815d81ded4be874bd\nReviewed-on: https://review.monogon.dev/c/monogon/+/162\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "30653eeb3adcbebfecb4216569dc247b51cfb464", "tree": "32171ac29f074a717c6e8ddb621a130f68428069", "parents": [ "fac8b2e265836dea105e8463a3a22b189764fd3f" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jun 17 15:44:29 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Mon Jun 21 09:12:58 2021 +0000" }, "message": "metropolis/proto/common: factor out common messages\n\nIn preparation for moving some proto structures from this proto package\n(eg. Node etcd serialization) into //metropolis/node/core/curator, we\nfully specify and document all elements of the Node/Cluster FSM states\nand a shared NodeRoles structure.\n\nThese will be shared by both serialized etcd state in the curator, the\ncurator API and possibly other APIs in the future. They are critical to\nthe Metropolis concept, so we aim to have them deduplicated across all\nuses, and that\u0027s why they\u0027re supposed to end up in\n//metropolis/proto/common.\n\nThese effectively temporarily duplicate definitions from\n//metropolis/proto/private, but these will be removed in an upcoming CL.\n\nTesting plan: no-op, in preparation for use in a further CL.\n\nChange-Id: I97087c7348a90e2dbe4cef03f7f55c7b58823ff1\nReviewed-on: https://review.monogon.dev/c/monogon/+/161\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "fac8b2e265836dea105e8463a3a22b189764fd3f", "tree": "0f78f138d0095d99a1bf529e29c29cb668a1f0b4", "parents": [ "b9013af7fa0247191099ec1f471a2d751537f545" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue May 04 12:23:26 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Tue Jun 08 10:23:53 2021 +0000" }, "message": "m/pkg/event: split out ValueWatch from Value\n\nSummary:\nThis implements a small TODO, letting the etcd Value implementation only\nimplement the Watch part of the interface.\n\nTest Plan: Refactor.\n\nChange-Id: I9ccd73ce4d165182d9588387230e71bcb425ab94\nReviewed-on: https://review.monogon.dev/c/monogon/+/122\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "b9013af7fa0247191099ec1f471a2d751537f545", "tree": "dec4a8c6e2ccefe41c2269f7938004a10d3792e1", "parents": [ "4166a71f51d9546c1dfd9f99b5fdffcb9301b57b" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Apr 29 16:47:56 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Tue Jun 08 10:23:53 2021 +0000" }, "message": "m/n/core/consensus/client: implement ThinClient()\n\nSummary:\nThis implements a ThinClient function on Namespaced which returns an\nclientv3.Client. These can be useful when dealing with existing code\nwhich expects this type, but only uses non-management APIs. For example,\nthe clientv3 concurrency library.\n\nTest Plan: To be used by future code, and basic enough that it IMO does not warrant unit tests?\n\nChange-Id: Ic6e38ff654bafd8bb385cf108637f8ed058015dc\nReviewed-on: https://review.monogon.dev/c/monogon/+/121\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "4166a71f51d9546c1dfd9f99b5fdffcb9301b57b", "tree": "a1e5341a9b71f973e1c24734872b0e1cc897f93c", "parents": [ "c89df2f0de65533e0801c6472cc4cee8b13cd761" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Mon Jun 07 21:58:54 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Tue Jun 08 00:02:06 2021 +0000" }, "message": "m/pkg/combinectx: implement\n\nThis implements combinectx, a Go library for combining two contexts into\na single one. We need this for the new curator logic (where we want to\ncancel RPC calls both when the incoming request gets canceled but also\nwhen leadership status changes), and this functionality has been\nfactored out as a reusable, generic library.\n\nPrior art:\n\n1) https://github.com/golang/go/issues/36503\n Proposal to add Merge() to context stdlib package. Unimplemented.\n\n2) https://github.com/teivah/onecontext\n Complex reflect-based logic for arbitrary amount of contexts to join,\n no functionality to detect which context caused the joined context to\n be canceled.\n\n3) https://github.com/LK4D4/joincontext\n No functionality to detect which context caused the joined context to\n be canceled.\n\nChange-Id: I774607da38b06c192ff0fee133eb258abd500864\nReviewed-on: https://review.monogon.dev/c/monogon/+/123\nReviewed-by: Leopold Schabel \u003cleo@nexantic.com\u003e\n" }, { "commit": "c89df2f0de65533e0801c6472cc4cee8b13cd761", "tree": "b65e4c12ab0c629dcc311335ad0151e1b19f3bbe", "parents": [ "dcf654592593e4ad897bfb34a5a9238a3223cca4" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Apr 27 15:51:37 2021 +0200" }, "committer": { "name": "Sergiusz Bazanski", "email": "serge@nexantic.com", "time": "Mon May 31 16:51:31 2021 +0000" }, "message": "m/pkg/event/etcd: implement etcd-backed Value\n\nThis implements Event Value stored in etcd, with each Value\ncorresponding to a single KV value stored in etcd.\n\nComes with more lines of unit tests than lines of code.\n\nChange-Id: I5514f211ded6640836ed801ddaf1b2fcc31ae552\nReviewed-on: https://review.monogon.dev/c/monogon/+/64\nReviewed-by: Lorenz Brun \u003clorenz@nexantic.com\u003e\n" }, { "commit": "216fe7b3ae949376467f626f339423a31ea7da97", "tree": "b0fe587b671a76bf6229339825d2a61df7fc847b", "parents": [ "6ebdc418f3c4799c12368e34ea78dc9c9757fb54" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri May 21 18:36:16 2021 +0200" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri May 28 17:54:03 2021 +0200" }, "message": "*: reflow comments to 80 characters\n\nThis reformats the entire Metropolis codebase to have comments no longer\nthan 80 characters, implementing CR/66.\n\nThis has been done half manually, as we don\u0027t have a good integration\nbetween commentwrap/Bazel, but that can be implemented if we decide to\ngo for this tool/limit.\n\nChange-Id: If1fff0b093ef806f5dc00551c11506e8290379d0\n" }, { "commit": "99d210d48afc2207ffb4064c58068faa9449a981", "tree": "781a73c0e5bf7e9ff586653eef0cce594b90def0", "parents": [ "4e0dba61375bcb989d86cacf18cf00ebfe6303b4" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Mon May 17 15:29:18 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue May 18 14:06:11 2021 +0200" }, "message": "m/n/k/plugins/kvmdevice: export resource name variable for easier consumption\n\nTrivial change to consume the resource name of this device plugin as\na variable.\n\nTest Plan: Trivial change\n\nX-Origin-Diff: phab/D791\nGitOrigin-RevId: d71d878f87be1da5a547e17b9965f92e737b644c\n" }, { "commit": "4e0dba61375bcb989d86cacf18cf00ebfe6303b4", "tree": "698e22e7ce446670c6833c60bf1bd58b5a1b5c65", "parents": [ "be57a039071a451763adc6c3456b7d79ca1999bb" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Mon May 17 15:25:15 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue May 18 14:06:08 2021 +0200" }, "message": "m/b/kube-code-generator: return go_generated_srcs output group for generated library\n\nTogether with D763 this allows the IntelliJ Bazel/Go code intelligence to pick up\nour generated Kubernetes clients. This sadly has some unrelated changes in there because\nbuildifier has opinions on code formatting.\n\nTest Plan: Tested in subsequent revisions.\n\nX-Origin-Diff: phab/D790\nGitOrigin-RevId: 7b0dc03cac63b5e79b3c2590614a4ed1f56a36ba\n" }, { "commit": "be57a039071a451763adc6c3456b7d79ca1999bb", "tree": "391ebab65e54c88c0b101a137371b283c5fd3812", "parents": [ "3536e4d4923e76486167c85c2b09a1cf4ca5502d" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue May 11 13:41:52 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue May 11 13:42:23 2021 +0200" }, "message": "m/test/launch: fix TPM tempdir permissions, wrap errors\n\nOn Linux, the following generally fails:\n\n $ cd /tmp\n $ mkdir test\n $ cd test/\n $ chmod 644 .\n $ touch foo\n touch: cannot touch \u0027foo\u0027: Permission denied\n\nThis changes our launch code to create a temporary TPM directory with\n755 instead of 644 permissions, preventing a situation like above\nmanifesting in our new CI.\n\nThis didn\u0027t manifest before as we always ran builds through podman, and\nthere this behaviour doesn\u0027t appear to hold, probably because we are uid\n0 there:\n\n $ podman exec -it monogon-dev bash\n bash-5.0# id\n uid\u003d0(root) gid\u003d0(root) groups\u003d0(root) context\u003dunconfined_u:system_r:spc_t:s0\n bash-5.0# cd /tmp/\n bash-5.0# mkdir test\n bash-5.0# cd test/\n bash-5.0# chmod 644 .\n bash-5.0# touch foo\n\nWe also drive-by some unwrapped error returns to be a bit more helpful.\n\nTest Plan: Tested on new CI, manually.\n\nX-Origin-Diff: phab/D773\nGitOrigin-RevId: 5a55a7878109717f0c17251a659dfc6ee04b94f4\n" }, { "commit": "68ca5eebd0ccd00a2d60eb42289c64357fb2e83f", "tree": "5706f5b4fa8dc44775dbabe24cd577f1d37a0422", "parents": [ "93bba15a0059da200a5d09a2bd7ec5ed5a667c60" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Apr 27 16:09:16 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue May 11 12:09:33 2021 +0200" }, "message": "m/pkg/event: move MemoryValue to subpackage\n\nThis keeps metropolis/pkg/event as a pure interface package, and\nmoves the memory-backed implementation to a subpackage.\n\nTest Plan: Refactor, coevered by tests.\n\nX-Origin-Diff: phab/D764\nGitOrigin-RevId: 1337bf55a7752293791b3efe8648bbf5f6e6e9e1\n" }, { "commit": "f055a7fce0263a30fd2c853b5ed002a765fc23e8", "tree": "de2dc0daeebfc7ecce2b1987ffb13eb4f2475088", "parents": [ "2666513457e8d7a282560a7090f35439ab9695ce" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Apr 13 16:22:33 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Wed Apr 14 14:35:09 2021 +0200" }, "message": "third_party/linux: build using unhermetic rule\n\nThis replaces ad-hoc genrules (for the node Linux image and the ktest\nimage) with a real Bazel rule with an attached transition which ensures\nwe end up with the same-ish configurations for all builds of an image.\n\nThis reduces rebuilds of the ktest Linux kernel, from three down to one.\n\nBefore: https://drive.google.com/file/d/1c6VmY2bqx9Pgs61TOUfgMi8Sn0WQeobu/view\n\nAfter: https://drive.google.com/file/d/13eO1rLhoBCMMRUKrmJz8QnhdAR3ctIGb/view\n\nWe also drive-by fix the Kubernetes CTS test suite to run on a single-node\nCluster (instead of failing early due to that being currently reworked).\n\nTest Plan: Build system refactor, following existing test.\n\nX-Origin-Diff: phab/D761\nGitOrigin-RevId: b5545ac5fd402fbf0340d941a90b9ea6ea0b6d43\n" }, { "commit": "2666513457e8d7a282560a7090f35439ab9695ce", "tree": "328d8f62ddb665b6cd057272f7cae2713aa247ad", "parents": [ "a105db57640d6abf6de368ec0c33a3a5b4f93893" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Apr 13 16:55:59 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue Apr 13 19:51:33 2021 +0200" }, "message": "Fix TCP BBR setting\n\nI previously set the TCP queuing discipline to BBR without actually compiling in BBR.\nSee T943. This actually builds in BBR and sets it as default in the kernel config, thus removing\nthe need to manually set it in userspace.\n\nTest Plan: CI\n\nBug: T943\n\nX-Origin-Diff: phab/D760\nGitOrigin-RevId: 779a709e4298ec59bfdcf462fe2f3563952204b6\n" }, { "commit": "a105db57640d6abf6de368ec0c33a3a5b4f93893", "tree": "dc96244d0b2922768d7f29431bab2a9ac581a1a6", "parents": [ "c75c9d4f30174c089cce47c608683bf09396fa30" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Mon Apr 12 19:57:46 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue Apr 13 19:51:13 2021 +0200" }, "message": "m/n/core/cluster: migrate to events and etcd namespaced client\n\nThis moves the status of the cluster manager to use a local event\nvariable. Watchers (like the node startup code) can now use this to get\nupdates on the state of the node and its cluster membership in a way\nthat\u0027s more abstracted from a sequential startup. This will permit us to\nmove a lof othe startup code into code common across different node\nlifecycle paths.\n\nTest Plan: Refactor, exercised by e2e.\n\nX-Origin-Diff: phab/D757\nGitOrigin-RevId: 31a3600ad2aab90a1e7f84d741e7ea40a0422724\n" }, { "commit": "c75c9d4f30174c089cce47c608683bf09396fa30", "tree": "a3d34c7ac52f16650de55ba28d3621add81e5f99", "parents": [ "886d2892d1717bc130cfa008742c06c29f7ff186" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Apr 13 16:40:14 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue Apr 13 19:51:11 2021 +0200" }, "message": "m/n/c/consensus: add client\n\nThis implementes a thin wrapper around etcd\u0027s clientv3.Client, with the\nfollowing advantages:\n\n - Only implements KV, Watcher and Lease interfaces, ie. unprivileged\n namespaceable interfaces - not cluster management interfaces. These\n will be available to both remote and local etcd connections.\n - Adds recursive namespacing functionality, which permits different\n parts of the subsystem to receive their own somewhat-sandboxed etcd\n subtree. This not only makes the etcd keyspace layout more strict,\n but also simplifies passing around etcd clients, as major components\n (like the kubernetes subsystem) can hand out its own sub-clients,\n instead of them having to be globally declared ahead of time.\n\nTest Plan: Exercised by tests.\n\nX-Origin-Diff: phab/D756\nGitOrigin-RevId: 03fead9a89c301a2e70df8a007b7ecb60b2364c7\n" }, { "commit": "886d2892d1717bc130cfa008742c06c29f7ff186", "tree": "e9d67752d55d52e948662521e96d83f343473dea", "parents": [ "0565ea3191d445bdaab647213d59ed3d7218630a" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Apr 13 16:39:39 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue Apr 13 19:51:08 2021 +0200" }, "message": "metropolis/test/launch: add InsecureKey\n\nCurrently, the node startup parameters are empty. Let\u0027s populate them with a development (\u0027insecure\u0027) key when started from the launch library.\n\nTest Plan: Future revision in stack will make use of this.\n\nX-Origin-Diff: phab/D754\nGitOrigin-RevId: 0cfa3c1d71911423ff169afc027edb768151de67\n" }, { "commit": "0565ea3191d445bdaab647213d59ed3d7218630a", "tree": "fd5f395dd2f54a1582538e7b5ac2b724db987625", "parents": [ "09c275bc489bc1de406be9a2e8f158eaa87b7c61" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Apr 13 11:52:00 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue Apr 13 19:51:06 2021 +0200" }, "message": "m/n/core/cluster: factor out bootstrap into separate file\n\nTest Plan: Refactor.\n\nX-Origin-Diff: phab/D753\nGitOrigin-RevId: 2e39db6673b2a0a6c1d7593f230fa691643e4c06\n" }, { "commit": "09c275bc489bc1de406be9a2e8f158eaa87b7c61", "tree": "31c62a93e37f6052aa99e2addacef6c060d75e85", "parents": [ "37050126ef89ec30cc677c272471debe55ec0d69" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Mar 30 12:47:09 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue Apr 13 11:04:05 2021 +0200" }, "message": "Add ftrace support to DebugService\n\nThis allows us to do ad-hoc kernel-level tracing on a running Metropolis node.\nUseful for tracking down complex bugs.\n\nExample: `bazel run //metropolis/cli/dbg -- trace -function_graph_filter blkdev_* function_graph`\n\nTest Plan: Debug utility, manually tested\n\nX-Origin-Diff: phab/D748\nGitOrigin-RevId: 924eb795250412a73eb30c0eef4a8c1cc726e5fd\n" }, { "commit": "37050126ef89ec30cc677c272471debe55ec0d69", "tree": "c64a64a622ec1c3e1e72fc12a6d4252c0e803cc1", "parents": [ "2999427c182463840a339cf0e82885d8a3b6e79f" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Mar 30 14:00:27 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue Apr 13 11:04:01 2021 +0200" }, "message": "Implement Block PVCs in our storage backend\n\nThis implements full support for Block PVCs in our Kubernetes storage backend.\nThe block PVCs are backed by files made available to the pods using loop devices and\nhave read-only and online expansion support.\n\nThis also requires a Kubernetes patch because they call losetup if block PVCs are used\nwith CSI to establish a form of lock on the backing block device. This lock is not\nexclusive and does absolutely nothing for our use case and could get very expensive\non dense machines so I removed it.\n\nTest Plan: Comes with E2E tests\n\nX-Origin-Diff: phab/D746\nGitOrigin-RevId: 430d3f445286c0d3498b2153df333a19f3fcab89\n" }, { "commit": "2999427c182463840a339cf0e82885d8a3b6e79f", "tree": "5e28c819a3db52d3fb77b92be6b78938a107d5cd", "parents": [ "30167f5cf55829d38f9d480466d7b5742c62a5fc" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Mon Apr 12 14:25:18 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue Apr 13 11:03:58 2021 +0200" }, "message": "metropolis/vm: add kube/{generated,apis}\n\nThis is a first pass at the vm.metropolis.monogon.dev Kubnernetes API,\ncontaining the VirtualMachine resource.\n\nThis also adds all the required BUILDfile boilerplate for the output\n(generated) directory.\n\nTest Plan: Builds should pass, eg. bazel build //metropolis/vm/kube/...\n\nX-Origin-Diff: phab/D752\nGitOrigin-RevId: d05c1a38e26b14c11b7a5d5885b15bbca8662174\n" }, { "commit": "30167f5cf55829d38f9d480466d7b5742c62a5fc", "tree": "fd89a3bb8a1c08b10d870a6b185b2deffa131cac", "parents": [ "9956e72c6c0b4f6436dc9493bc213965ee0cc191" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Wed Mar 17 17:49:01 2021 +0100" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue Apr 13 11:03:56 2021 +0200" }, "message": "Add VM infrastructure smoke test\n\nThis adds an E2E test which exercises the VM infrastructure (Kubernetes, KVM device plugin and QEMU).\nThis test should ensure that nobody breaks the core infrastructure Metropolis VMs rely on.\n\nTest Plan: This is a test\n\nX-Origin-Diff: phab/D740\nGitOrigin-RevId: ddf629725dfb664ace5a50efee9ed9442962d6f7\n" }, { "commit": "9956e72c6c0b4f6436dc9493bc213965ee0cc191", "tree": "7842ac67432e3a187dda6a2dcb46d11088934159", "parents": [ "dca59d924dac4345099e5acd99405b5451d29cdb" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Wed Mar 24 18:48:55 2021 +0100" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Tue Apr 13 11:03:53 2021 +0200" }, "message": "Add Loop Device package\n\nThis adds Loop device support in our Linux kernel and adds a Go package for working with them.\nIt also drive-by adds a pre-mounted tmpfs to ktest as that is quite useful in a lot of situations.\n\nTest Plan: Comes with ktests.\n\nX-Origin-Diff: phab/D745\nGitOrigin-RevId: fa06bcdddc033efb136f56da3b4a91159273bf88\n" }, { "commit": "7353e17fcd3b158e575ca4455ee75224fa4f0921", "tree": "7208b1e0dd5ba9af5f97c31a49aa5007525bc4c3", "parents": [ "b76580c0189209f7feef90f0900e0758aaf122ce" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Mar 31 22:09:22 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Thu Apr 01 18:58:58 2021 +0200" }, "message": "metropolis/build: add kube-code-generator\n\nThis implements kube-code-generator, a set of Bazel rules for generating\nKubernetes resource APIs based on a Go library, using\nk8s.io/code-generator.\n\nTest Plan: Was considering adding a test for this - but this is practically best tested with the next change in the stack, which actually uses it to implement the VM hypervisor kube API.\n\nX-Origin-Diff: phab/D751\nGitOrigin-RevId: 31e3b632c2e83282c8b2c415402cddea66d4ce51\n" }, { "commit": "b76580c0189209f7feef90f0900e0758aaf122ce", "tree": "72cd514a620bb623ca7056a9b21e158bbc89b4fd", "parents": [ "339f97dc7ae48876f77b1195a8840f9369fb5d25" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Mar 31 22:07:01 2021 +0200" }, "committer": { "name": "Leopold Schabel", "email": "leo@nexantic.com", "time": "Thu Apr 01 18:58:55 2021 +0200" }, "message": "metropolis/build: add gotoolwrap\n\nThis adds gotoolwrap, a tiny Go executable used to wrap binaries that\nwant to access the monogon workspace as a GOPATH during build steps.\n\nTest Plan: Used further down the stack in code generation.\n\nX-Origin-Diff: phab/D750\nGitOrigin-RevId: 83d11b94d025d3652fce88917b1664d93454c60f\n" }, { "commit": "c4a3aab4c8d14d05b7e2448ab11897b6f9093046", "tree": "61585da074852ab286fab9fefa671e35fc1e4658", "parents": [ "42e61c6a3601d7f63176af3f53a3ac9c53120099" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Mar 16 13:31:19 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Mar 16 13:31:19 2021 +0100" }, "message": "metropolis/proto: remove internal\n\nThis is now unused (as it has been replaced by proto/private) and can be\nremoved.\n\nTest Plan: Refactor.\n\nX-Origin-Diff: phab/D736\nGitOrigin-RevId: e71dfd70d36d0ae9b29e2b287fcc337cf2520a51\n" }, { "commit": "42e61c6a3601d7f63176af3f53a3ac9c53120099", "tree": "2fd6e86b1a0dee6b4928f04ac359b4b262806d73", "parents": [ "34fe8c666811c63f33ba970ba0e39d25804ab4a3" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Mar 18 15:07:18 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Mar 18 15:07:18 2021 +0100" }, "message": "m/n/core/cluster: rewrite bootstrap to conform to cluster lifecycle DD\n\nThis removes the existing cluster/manager code and reimplements it from\nscratch, finally implementing the cluster lifecycle design document for\ncluster bootstrap.\n\nTest Plan:\nE2e should cover this. Maybe we could unit test the manager? But that would\nrequire a ton of DI work. Not sure if it\u0027s worth it.\n\nX-Origin-Diff: phab/D735\nGitOrigin-RevId: b00c97b0a102a21605d16086df82a6ece6eb7f4d\n" }, { "commit": "34fe8c666811c63f33ba970ba0e39d25804ab4a3", "tree": "fce90696e07eb76efea1436047393ad66ac04404", "parents": [ "3ea1a3a30d41980a8a5b5f321b3e7bc8af96dc29" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Mar 16 13:20:09 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Mar 16 13:20:09 2021 +0100" }, "message": "m/node/core/consensus: always use member name as address\n\nInstead of explicitly giving the consensus service an address to listen\non, we now just give it a hostname. Etcd will resolve the name itself on\nstartup. It\u0027s not yet known if it will re-resolve it and handle changes.\nIf that\u0027s not the case, we will have to implement a restarting\nbabysitter process instead.\n\nTest Plan: E2e tests should cover this.\n\nX-Origin-Diff: phab/D734\nGitOrigin-RevId: c2255b2956137e2bbd705ac1965418c0540d0046\n" }, { "commit": "3ea1a3a30d41980a8a5b5f321b3e7bc8af96dc29", "tree": "0d4ab65ff6af76e3115ac8a82452216301aebce4", "parents": [ "72068da814af80568cb106b877ef8f5e526e684c" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Mar 16 13:17:33 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Mar 16 13:17:33 2021 +0100" }, "message": "m/node/core/conensus: handle etcd restarts\n\nThis makes the etcd service more reliable in case of transient failures\nwhen starting in a new cluster. Previously, any restart of etcd on the\nfirst (bootstrapping) node would cause etcd to get stuck and never start\nagain (as certificates were already created). This changes the logic to\nallow existing certificates.\n\nThis also handles the case of etcd attempting to start as the network is\nreconfigured, and eg. the external hostname is not yet resolvable.\n\nTest Plan:\nNo tests yet. This should be tested by a more comprehensive e2e test where we\nrandomly kill some runnables (see: T872).\n\nX-Origin-Diff: phab/D733\nGitOrigin-RevId: 8ac426f9423ec2353537eec651071e99a5e5ec53\n" }, { "commit": "72068da814af80568cb106b877ef8f5e526e684c", "tree": "31831657903e62d4a813c9d18d47413154f13560", "parents": [ "d8af5bf4c14a5c53d1736695a1210b6eea4d246e" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Mar 16 13:15:45 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Mar 16 13:15:45 2021 +0100" }, "message": "metropolis/proto: add private\n\nWe add a new proto package, \u0027private\u0027. This will hold all non-public\nproto messages for the implementation of the cluster lifecyclce design\ndocument.\n\nWe duplicate \u0027internal\u0027 for two reasons:\n\n - make it easier to port code accross to use the new protos/cluster\n code, while slowly phasing out protos/internal.\n - \u0027internal\u0027 was a poor name choice, as it\u0027s significant in Go path\n naming, and might cause some problems in the future. We have Bazel\u0027s\n visiblility mechanism to protect accidental use.\n\nTest Plan: New proto, not yet used - see rest of stack.\n\nX-Origin-Diff: phab/D732\nGitOrigin-RevId: 15632353c46068b4f4c5025828c9c8459feaa335\n" }, { "commit": "d8af5bf4c14a5c53d1736695a1210b6eea4d246e", "tree": "83e4c35e105d8fb0e5bf71ac8c93bbd3e7f964b4", "parents": [ "4e090357c4f1f3bae53a5f2feaf20ea5e1bbbe61" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Mar 16 13:38:29 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Tue Mar 16 13:38:29 2021 +0100" }, "message": "metropolis/node: use Event Value for network status\n\nThis moves over the GetIP API to use our fancy new event/value library.\nThe consumers of this data are currently the cluster manager and the\nkubernetes root service. Both are migrated over.\n\nTest Plan: Refactor, covered by E2E tests.\n\nX-Origin-Diff: phab/D711\nGitOrigin-RevId: 8a1e0dd35236d55492722f4439323cb2ee9574fc\n" }, { "commit": "4e090357c4f1f3bae53a5f2feaf20ea5e1bbbe61", "tree": "335ec273335722befdeca623b8f3f787a2cd6571", "parents": [ "0ed2f96a3a86aff2c9ce36289aa5d58a75f4d59b" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Wed Mar 17 17:44:41 2021 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Wed Mar 17 17:44:41 2021 +0100" }, "message": "Add KVM device plugin\n\nThis adds a KVM device plugin for Kubernetes. This plugin allows for unprivileged access and granular\ncontrol of KVM access.\n\nTest Plan: Tested in subsequent revision\n\nX-Origin-Diff: phab/D739\nGitOrigin-RevId: 5cd738a47d24e7bfdc29bbd1a31537209e1ebf46\n" }, { "commit": "0ed2f96a3a86aff2c9ce36289aa5d58a75f4d59b", "tree": "afbe1fb6cd0a1667e981edfe97969338437bdaca", "parents": [ "056042962060369bd7607ecfea51c515fc3a8140" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Mon Mar 15 16:39:30 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Mon Mar 15 16:39:30 2021 +0100" }, "message": "metropolis/proto: EnrolmentConfig -\u003e NodeParameters\n\nThis starts off the move to a node configuration API conforming to\nthe lifecycle management design document.\n\nInstead of an Enrolment Config used only to join an existing cluster, we\nmove to a NodeParameters proto that must always be given to a node if\nit\u0027s supposed to either bootstrap a new cluster or join an existing one.\n\nThis links the existing cluster management code (and its state machine)\nto work with this file. However, that state machine will be removed very\nsoon, anyway.\n\nWe also remove everything related to golden tickets.\n\nThis breaks multi-node tests.\n\nX-Origin-Diff: phab/D710\nGitOrigin-RevId: f22615fbccab975f2d5e6928bdc7387ab3aa5714\n" }, { "commit": "056042962060369bd7607ecfea51c515fc3a8140", "tree": "86a6dbf7b1781ed2f5baf332938d4e8211353112", "parents": [ "0ab4edafde3eb22e111e75d6aa5e29faa92c30ca" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri Mar 12 17:47:21 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri Mar 12 17:47:21 2021 +0100" }, "message": "m/node/kubernetes: parse klog output from services\n\nThis translates Kubernetes\u0027 logging ingo logging that we can\nquery/filter more easily.\n\nTest Plan: We don\u0027t test resulting logs from the system, and I\u0027m not sure we should?\n\nX-Origin-Diff: phab/D716\nGitOrigin-RevId: ba3f42b9a4e3172bf058bd7dce4283f50dc8e69d\n" }, { "commit": "0ab4edafde3eb22e111e75d6aa5e29faa92c30ca", "tree": "8931f10cd69309ece470c38c3a062ef74f3699a5", "parents": [ "9411f7c2ed0afbbf617075ab37901addc76fadfb" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri Mar 12 17:43:57 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri Mar 12 17:43:57 2021 +0100" }, "message": "m/pkg/logtree: implement klog parsing\n\nThis adds logtree.KLogParser, a shim which parses klog/glog-formatted\nlines into logtree leveled logging.\n\nThis will be used to consume logs from external components (like\nKubernetes services) into leveled logging inside logtree.\n\nAn alternative would be to switch all Kubernetes components to\n\u0027structured\u0027 (JSON) logging - but that seems to still be experimental,\nand does not exactly map into something that we can log further. Maybe\nin the future we can switch over, and also copy these over into our own\nbinary/structured logging.\n\nTest Plan: Adds unit tests for parsing, which is the most tricky part.\n\nX-Origin-Diff: phab/D715\nGitOrigin-RevId: 9994d819f15c9542800d488f57c83ab945a35d34\n" }, { "commit": "9411f7c2ed0afbbf617075ab37901addc76fadfb", "tree": "f1f62aa538ba3c2265815d2dbe942377264850a5", "parents": [ "0de189355c6afad6f677029d90fa40dee824141b" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Mar 10 13:12:53 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Mar 10 13:12:53 2021 +0100" }, "message": "m/node/kubernetes/pki: refactor out CA functionality\n\nThis factors out all non-k8s-specific CA functionality from\nmetropolis/node/kubernetes/pki into metropolis/pkg/pki.\n\nThis will allow us to re-use the same PKI-in-CA system to issue\ncertificates for the Metropolis cluster and nodes.\n\nWe also drive-by change some Kubernetes/PKI interactions to make things\ncleaner. Notably, this implements Certificate.Mount to return a\nfileargs.FileArgs containing all the files neede to use this\nCertificate.\n\nTest Plan: covered by current e2e tests. An etcd harness to test this independently would be nice, though.\n\nX-Origin-Diff: phab/D709\nGitOrigin-RevId: bdc9ff215b94c9192f65c6da8935fe2818fd14ad\n" }, { "commit": "ddd6caff9edac56dad727a79eb5b0faf4dbd6cb9", "tree": "120710eb4a9acf0c3ad1086d9f6f6f3c850a0d70", "parents": [ "bcae658f9530e95cde2ac931beacae71c9fb240e" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Thu Mar 04 17:16:04 2021 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Thu Mar 04 17:16:04 2021 +0100" }, "message": "Build mkfs.xfs using rules_cc\n\nThis drops the old big genrule for mkfs.xfs and replaces it with a nice rules_cc build system\nwith the help of bazel_cc_fix generated patches and our musl-based toolchain.\nWhile we\u0027re at it I bumped the versions of all related dependencies to their latest stable release.\nThis also means pulling in ini.h which is a dependency of the new xfstools version.\n\nInstructions to regenerate the patches are included in the spec files.\n\nToolchain selection is done by the existing transition in our rootfs rule so we automatically get a musl-built\nstatic binary when building for the rootfs.\n\nTest Plan: Tested with E2E tests, should fail fairly catastrophically if something were wrong.\n\nX-Origin-Diff: phab/D708\nGitOrigin-RevId: 648a05cdd08cfa84a8a9f4c057c52446e7005631\n" }, { "commit": "c00318e448212b01a8121059be3c3e9b35bd13a7", "tree": "38011616a2112e14591da1b06ac65ac8ec75b372", "parents": [ "32d73486f4ea778cd3ea58e2d579e862cf67fb9c" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Mar 03 12:39:24 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Wed Mar 03 12:39:24 2021 +0100" }, "message": "m/pkg/event: implement\n\nThis specifies event.{Value,Watcher}, an interface for data that might\nbe updated by its producer, and which is watched for such updates by\nmultiple consumers.\n\nIt also implements MemoryValue, a Value that is stored in memory.\n\nTest Plan: adds unit tests.\n\nX-Origin-Diff: phab/D706\nGitOrigin-RevId: 271fd4e88969817b66318d3e03d50b70cf2819b8\n" }, { "commit": "32d73486f4ea778cd3ea58e2d579e862cf67fb9c", "tree": "78e3444e0b55df55f512415dbfd34977cdca2350", "parents": [ "6c4199afe4dc4d446679b862f528e840e60925df" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Mon Feb 01 23:49:17 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Mon Feb 01 23:49:17 2021 +0100" }, "message": "metropolis: introduce AAA.Escrow RPC\n\nThis is a combined proto change and design document RFC.\n\nThis implements a generic \u0027Escrow\u0027 methid, used to allow external\nentities to log into a Metropolis cluster. This flow\u0027s subject vaguely\ncorresponds to \u0027Entity\u0027 objects from the Lifecycle DD, but this will be\nmore precisely defined in a subsequent change which introduces the\nactual entities objects, the way they\u0027re identified, and the way they\u0027re\nstored in the cluster.\n\nIn addition, this formalizes the part of the LDD in which entities are\nable to perform hardware attestation on nodes. The hardware attestation\npart is not fully implemented, but is placed within the bounds of the\nEscrow streaming RPC. Entities might also be able to performs this\nhardware attestation in a separate RPC call (having already requested a\nshort-lived certificate permitting access to RPC), but this is not yet\nsure.\n\nThis design, is in a way, a modernized version of GSSAPI. It assumes it\nruns over a confidential channel (TLS), and that it only ever returns\nx509 certificates emitted for the requesting client. It is also designed\nto handle flows that we expect to use within Metropolis.\n\nThis design has some known limitations:\n\n1) Limited decisionmaking abitility by the server to decide which proofs\n are needed - ie., the server cannot change its mind what other proofs\n are needed as the client presents some. Currently the server can\n decide the proofs only based on the parameters given by the client,\n and the initial context of the connection, ie. its originating\n address and the presented TLS certificate.\n2) Limited expressibility of required proofs to the client, currently\n all listed must be fulfilled.\n\nThis, however, can be extended as the protocol evolves, and can continue\nto support simple clients that handle only this protocol. Especially 2)\nmight be limiting us from preventing things like accepting emergency\ncertificates without necessarily needing an OIDC login, even though OIDC\nlogins are required for other kinds of certificates. We are explicitly\ntrying to keep things simple for now, and just not write ourselves into\na corner here.\n\nFinally, this API should cover all scenarios expressed within T865 -\nminus the entity storage part within the cluster.\n\nTest Plan: Proto change and review process.\n\nX-Origin-Diff: phab/D698\nGitOrigin-RevId: 92892b5522a4d41d572fd4c10f24d26f72919aeb\n" }, { "commit": "6c4199afe4dc4d446679b862f528e840e60925df", "tree": "fcea4d9f54f2e04cf1d203e104c4a14bfa103702", "parents": [ "5999e92b2da34cbbd50391327ec01081a91866ee" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Wed Feb 10 17:34:29 2021 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Wed Feb 10 17:34:29 2021 +0100" }, "message": "Set reasonable defaults for our kernel\u0027s network configuration\n\nThis sets a number of sysctl options to tune the kernel for a datacenter-like environment by increasing\nbuffers and choosing a better congestion control algorithm. It also enforces reverse path filtering to\nprevent spoofing from CAP_NET_ADMIN-enabled containers and blocks source routing as we have no need for that\nand it might some day interfere with policy efforts.\n\nTo set all these options a small helper structure has been added which makes setting these more compact\nand nicer to read.\n\nTest Plan: Covered by E2E for breakage, scalability improvements not yet testable\n\nBug: T495\n\nX-Origin-Diff: phab/D704\nGitOrigin-RevId: 427b2513d604090e51b37587d772f240112be09d\n" }, { "commit": "5999e92b2da34cbbd50391327ec01081a91866ee", "tree": "164e447b7d17e89f2b1046c3da51af141deaa08b", "parents": [ "3a99c590543394ceb5260282ef8e924b44e8eef8" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Wed Jan 27 18:53:54 2021 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Wed Jan 27 18:53:54 2021 +0100" }, "message": "Bump Linux kernel to 5.10\n\nThis bumps our Linux kernel to 5.10. There\u0027s one minor fix in fsinfo accounting for the fact that strings are\nnow null-terminated. While debugging this I also drive-by fixed a minor typing issue in quotactl.go.\n\nThis drops support for the old initramfs loading method (which was the driving force for the EROFS changes)\nas refactors in the kernel made the patch we carried until now non-viable. Nothing uses it anymore, everything is\neither a microvm-style machine which doesn\u0027t use EFI and thus doesn\u0027t suffer from the issue or uses EROFS.\n\nTest Plan: No new functionality, should be covered by E2E tests.\n\nX-Origin-Diff: phab/D697\nGitOrigin-RevId: d8e40954abb66cb082eecbca372b94a7e40b84a8\n" }, { "commit": "3a99c590543394ceb5260282ef8e924b44e8eef8", "tree": "e1b727a0c12b387e1bc12d71826405b8b588fa40", "parents": [ "6b13bf1a98c4a612d13ae939e68802e77fb45474" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jan 26 19:57:21 2021 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jan 26 19:57:21 2021 +0100" }, "message": "Switch Metropolis to EROFS-based root filesystem\n\nThis gets rid of the old large initramfs and switches to an EROFS-based root\nfilesystem. It also drops the copy \u0026 remount compatibility code. As this filesystem is\nproperly read-only and not just ephemeral, this also brings various changes to the code\nto make systems compatible with that.\n\nTest Plan: Covered by E2E tests, also manually smoke-tested.\n\nX-Origin-Diff: phab/D696\nGitOrigin-RevId: 037f2b8253e7cff8435cc79771fad05f53670ff0\n" }, { "commit": "6b13bf1a98c4a612d13ae939e68802e77fb45474", "tree": "bcd1f6566df193d026889da174a6773b8689d643", "parents": [ "10b9ee96d4c2b8a011af4cd4db3390c1fd1ddf93" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jan 26 19:54:24 2021 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jan 26 19:54:24 2021 +0100" }, "message": "Add EROFS creation utility and Bazel rule\n\nThis adds a binary which can create EROFS filesystems from a spec and a Bazel rule similar to\nnode_initramfs which creates EROFS filesystems.\n\nTest Plan: Tested in subsequent revision\n\nX-Origin-Diff: phab/D695\nGitOrigin-RevId: 4e8f970938c93606da46f475387de1e013b1c35c\n" }, { "commit": "10b9ee96d4c2b8a011af4cd4db3390c1fd1ddf93", "tree": "ea0ca7da66e44cc52defa1a307e47642ca83a150", "parents": [ "2073ce34e57b0be3cedd39b8934869abb6f73582" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Wed Feb 10 12:14:23 2021 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Wed Feb 10 12:14:23 2021 +0100" }, "message": "erofs: Don\u0027t modify caller\u0027s data\n\nThe erofs library\u0027s directory writer appends data to parameters. Because of the way slices work in Go this\nresults in the caller\u0027s data being changed, which is obviously undesirable. Fix this by making a copy first.\n\nTest Plan: Minimal change, should be covered by existing tests\n\nX-Origin-Diff: phab/D703\nGitOrigin-RevId: ebf473c1049e5e8035802382220aba98c4498877\n" }, { "commit": "378a4455aedda838f60c546e55199092f24952ed", "tree": "aa78b858535224fe8c9b24c2ff7e9ed2c903080b", "parents": [ "74e8e5c35fea1ec9ce13c8a2d16100bab45d42d9" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jan 26 13:47:41 2021 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jan 26 13:47:41 2021 +0100" }, "message": "Add EROFS library\n\nThis adds a library to write EROFS filesystems. It supports most of the non-deprecated features the\nfilesystem supports other than extended inodes (which have no benefits for most use cases where EROFS would be\nappropriate). EROFS\u0027s variable-length extent compression is partially implemented but it requires an LZ4\ncompressor with support for fixed-size output which Go\u0027s https://github.com/pierrec/lz4 doesn\u0027t have. This means\nthat VLE compression is currently not wired up.\n\nThis will be used later as a replacement for our current initramfs-based root filesystem.\n\nTest Plan: Has both integration and some unit tests. Confirmed working for our whole rootfs.\n\nX-Origin-Diff: phab/D692\nGitOrigin-RevId: 8c52b45ea05c617c80047e99c04c2b63e1b60c7c\n" }, { "commit": "74e8e5c35fea1ec9ce13c8a2d16100bab45d42d9", "tree": "3ec734c4b86fed54a5039623c789dd4b805b3b6e", "parents": [ "19eb0006edc79edc53fb53ea0eed67e93f4c8eba" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jan 26 14:00:50 2021 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jan 26 14:00:50 2021 +0100" }, "message": "Make containerd work with read-only root\n\nThis makes containerd work with a read-only root. There were a few config mistakes on our side which\ncaused it to write to the rootfs (mostly leftovers from the switch to /ephemeral) and a semi-hardcoded path\nin /var/lib/cni from containernetworking/cni. This is technically configurable, but it would require patching\nthree different repos (see diff message) and getting all of them to agree to take the change and wait for\nit to propagate to all repos (containerd is known to be slow to release stuff). So let\u0027s just hack in\nthis one-line diff for the time being.\n\nTest Plan: Should be covered by existing tests\n\nX-Origin-Diff: phab/D694\nGitOrigin-RevId: 0e8f5dbfb216539c16e64130af9fe1023722ae1b\n" }, { "commit": "19eb0006edc79edc53fb53ea0eed67e93f4c8eba", "tree": "704a52ab75bde43409d80246cf23bce6b6be3467", "parents": [ "842536b10bd1b11e62317940feef215442a8ecb4" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jan 21 14:25:25 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jan 21 14:25:25 2021 +0100" }, "message": "third_party: bump Kubernetes to 1.19.7\n\nThis... didn\u0027t exactly go well. Turns out a change between rc.1 and rc.2\nbroke our runc runtime by enabling seccomp by default for pod sandboxes.\n\nWe work around this by reverting this change, and filing T916 to solve\nthis soon.\n\nThis fixes T910 and T909.\n\nTest Plan: kube bump, CI should run e2e, didn\u0027t run CTS.\n\nBug: T910, T909\n\nX-Origin-Diff: phab/D691\nGitOrigin-RevId: 78afca77c294895859e0af9150128d82677d875b\n" }, { "commit": "842536b10bd1b11e62317940feef215442a8ecb4", "tree": "264906157b5cd51ef39e952326b85da48b5bbb5b", "parents": [ "f12bedfa4cd144c3abc4deac58405067d55f9c87" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jan 26 13:54:57 2021 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Jan 26 13:54:57 2021 +0100" }, "message": "Make Kubernetes work with read-only root\n\nThis makes Kubernetes work with a read-only root. There\u0027s two places where they hardcode\npaths: One is the DeviceManager socket path (/var/lib/kubelet/device-plugins/kubelet.sock), that one\nis easy to fix because KubeletRootDir is available one scope above. The other one is the pod logs dir\nwhich is too far removed from the main Kubelet config, so I just changed their hardcoded path to ours.\nThe first patch should be upstreamable, for the second one we\u0027d need to take a different approach to upstream.\n\nTest Plan: Should be covered by existing E2E tests.\n\nX-Origin-Diff: phab/D693\nGitOrigin-RevId: 4606ab228a24bd4a0274f8e3156123710a59f2aa\n" }, { "commit": "f12bedfa4cd144c3abc4deac58405067d55f9c87", "tree": "ddbc408e424a0ea8e446bcf0022ee16278202d63", "parents": [ "c3ad846e0eaf4cf008130a643ff247aa27531e17" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri Jan 15 16:58:50 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri Jan 15 16:58:50 2021 +0100" }, "message": "*: bump up Go dependencies\n\nThis started off as \u0027let\u0027s bump gVisor\u0027. However, pulling that thread\nresulted in quite a few things that also required bumping for the build\nto actually work. Here I come back from a day in the Bazel mines,\nbearing fruits of my labor.\n\nNotable changes:\n\n - bump up gVisor\n - bump up containerd\n - bump up Bazel\n - bump up rules_go, rules_docker, Gazelle\n - use google.golang.org/protobuf (the \u0027new\u0027 go proto package)\n - bump up gRPC (but not too much, as go-etcd is still straggling)\n\nNotable effects:\n\n - new gVisor supports TTY allocation (kubectl run -it\n --image\u003dubuntu:20.04 ubuntu bash now works!)\n\nNotable notes:\n\n - gVisor shim has new been rolled into the main gVisor package and is\n slightly easier to build (we can get rid of a bunch of patches).\n - Opencontainers\u0027 runtime-specs now follow containerd instead of gVisor\n - gVisor had to be taught to use the slightly newer runtime-specs via a\n new patch.\n - go_rule() in Starlark is now deprecated, and we had to change our\n Starlark rule definitions to use rule() instead. We also had to patch\n gVisor to do that (as there hasn\u0027t yet been a release that rolled\n this up).\n - Gazelle now supports different naming schemes for generated Go\n targets - either the old //foo/bar:go_default_library scheme, or a\n new and nicer //foo/bar:bar scheme. We currently force the usage of\n the old scheme, as switching over is probably not going to be easy\n (we use a lot of external Bazel files, and we have to wait for their\n compatibility with the new scheme first).\n - New Bazel/rules_go sets a TMPDIR long enough to generate paths (via\n ioutil.TempDir) to which sockets cannot be bound (108-byte limit).\n - The new protobuf API is incompatible with gogoproto. containerd/ttrpc\n uses gogoproto, but we are smart enough to pull in the old protobuf\n library as gogoproto\u0027s transitive dep. However, ttrpc also wants to\n use some proto-generated grpc bits, and that doesn\u0027t work. We have to\n pull in a ttrpc fork from a PR that hasn\u0027t yet been merged that fixes\n this issue.\n\nTest Plan: Refactor only, should be covered by tests.\n\nX-Origin-Diff: phab/D689\nGitOrigin-RevId: 1188c0605d25e7f40307fab5fd96e7019f3a9171\n" }, { "commit": "c3ad846e0eaf4cf008130a643ff247aa27531e17", "tree": "d9a8287f49f596c8ffd9c7d0dce6939c16a06707", "parents": [ "febf0b0defa3a4ccfb58f4c77ddb095a2668162a" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri Jan 08 16:45:51 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Fri Jan 08 16:45:51 2021 +0100" }, "message": "ci: build both debug and non-debug builds, add secondary cache\n\nThis runs `test //... -c dbg` in addition to `test //...`. Because switching to-and-from the debug configuration causes the local cache to get thrashed, we add a secondary cache via --disk_cache. This should, at some point, be replaced with a proper remote cache instead.\n\nWe also drive-by fix a debug build issue.\n\nFixes T883.\n\nTest Plan: This should test more things in CI now.\n\nBug: T883\n\nX-Origin-Diff: phab/D688\nGitOrigin-RevId: 9c35c4737d6b205a6bc74d50665c37535ac1d5ba\n" }, { "commit": "31370b07f0df2dc2765d812d4ce00a6b35185b16", "tree": "15563902eee9591083284441c8505b084b275d0a", "parents": [ "313816f41244d7520eb2b6f8c231328ee5b7a4ef" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jan 07 16:31:14 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jan 07 16:31:14 2021 +0100" }, "message": "*: git.monogon.dev -\u003e source.monogon.dev\n\nThis implements T882, setting our (virtual) GOPATH to source.monogon.dev\nfor this repository.\n\nTest Plan: Refactor, CI only.\n\nX-Origin-Diff: phab/D686\nGitOrigin-RevId: c5e2309089948ffc3a98e68e2e0e1cbb157d3a36\n" }, { "commit": "313816f41244d7520eb2b6f8c231328ee5b7a4ef", "tree": "7a8aa08b48b22a7843afd76c14b8fd2ebe3c2cfb", "parents": [ "d9ed6560d5db2783252945e1dd3c2e4f908c019e" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Dec 22 16:52:26 2020 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Dec 22 16:52:26 2020 +0100" }, "message": "Disable integrity and encryption for debug builds\n\nThis disables disk encryption and integrity when running in the debug profile.\nIt also makes mkfs.xfs not issue discards as the backend is either a sparse file\n(during development) or dm-crypt/dm-integrity which ignores them. In both cases\nthey are counterproductive and slow things down.\n\nTest Plan: Not exercised by normal tests, manually tested.\n\nX-Origin-Diff: phab/D681\nGitOrigin-RevId: 24fae1e3de8d852b414ebd50f3fbe032440683fb\n" }, { "commit": "0be9be88224dd87eedb10436b11615fa59862271", "tree": "2cffcd0ca273ada48c0b42a36bd25bb1cc2da35c", "parents": [ "549b72b2d65051403301f53111509f77e88b379b" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jan 07 15:23:44 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jan 07 15:23:44 2021 +0100" }, "message": "metropolis: Lock down visibility rules\n\nThis formalizes the package structure introduced by D683.\n\nTest Plan: Pure refactor, CI only.\n\nX-Origin-Diff: phab/D684\nGitOrigin-RevId: 574aa14c71faf94f4a5c02a2110e2e3fef7d36ac\n" }, { "commit": "549b72b2d65051403301f53111509f77e88b379b", "tree": "b4e523d5d17e8130545e58b58870b4a18118a780", "parents": [ "696f39abb19ffcca03e9fc5a98681338216b1e7f" ], "author": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jan 07 14:54:19 2021 +0100" }, "committer": { "name": "Serge Bazanski", "email": "serge@nexantic.com", "time": "Thu Jan 07 14:54:19 2021 +0100" }, "message": "metropolis: unify utility packages\n\nOne last sweeping rename / reshuffle.\n\nWe get rid of //metropolis/node/common and //golibs, unifying them into\na single //metropolis/pkg meta-package.\n\nThis is to be documented somwhere properly, but here\u0027s the new logic\nbehind selecting where to place a new library package:\n\n - if it\u0027s specific to k8s-on-metropolis, put it in\n //metropolis/node/kubernetes/*. This is a self-contained tree that\n other paths cannot import from.\n - if it\u0027s a big new subsystem of the metropolis core, put it in\n //metropolis/node/core. This can be imported by anything in\n //m/n (eg the Kubernetes code at //m/n/kubernetes\n - otherwise, treat it as generic library that\u0027s part of the metropolis\n project, and put it in //metropolis/pkg. This can be imported by\n anything within //metropolis.\n\nThis will be followed up by a diff that updates visibility rules.\n\nTest Plan: Pure refactor, CI only.\n\nX-Origin-Diff: phab/D683\nGitOrigin-RevId: 883e7f09a7d22d64e966d07bbe839454ed081c79\n" }, { "commit": "696f39abb19ffcca03e9fc5a98681338216b1e7f", "tree": "3eb962a59b2af0c3fb3cf40a05c405ae23b7f8a9", "parents": [ "81de89b8675ee0ce677225ffef1cd3ee6ad9f56f" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Dec 22 16:54:01 2020 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Tue Dec 22 16:54:01 2020 +0100" }, "message": "Make kube-apiserver use supervisor helper for commands\n\nAll other Kubernetes services already use this and it enables them to log\ninto logtree. Make kube-apiserver also do the same.\n\nTest Plan: Covered by existing tests\n\nX-Origin-Diff: phab/D680\nGitOrigin-RevId: 59df1342edc2cb27c22ffa9b4eb9101d7d1b400f\n" }, { "commit": "6df7c4f6b2c9a896357cb6c4e236d588f4e23277", "tree": "1f6690534fa123509708fe197ef3f35e6edf6903", "parents": [ "df952416e693f5b3180f1e69b6021a589cdc80d9" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Mon Dec 21 15:02:00 2020 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Mon Dec 21 15:02:00 2020 +0100" }, "message": "Properly exit CTS on signals\n\nThis kills the loops when the global context is cancelled ensuring reliable termination\nof the CTS with Ctrl+C or a signal.\n\nTest Plan: Manually tested by aborting the CTS with Ctrl+C\n\nX-Origin-Diff: phab/D673\nGitOrigin-RevId: a2f367cd4a7a57bb573bd57656148681add048a2\n" }, { "commit": "df952416e693f5b3180f1e69b6021a589cdc80d9", "tree": "9ad8fdce45516e723b3ebe1aa49aec8b2d8194e6", "parents": [ "662b5b3119b0798980b887d1ef9fa1b5632aa7fb" ], "author": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Mon Dec 21 14:59:36 2020 +0100" }, "committer": { "name": "Lorenz Brun", "email": "lorenz@nexantic.com", "time": "Mon Dec 21 14:59:36 2020 +0100" }, "message": "Fix nanoswitch logging after logtree refactor\n\nThe LogTree refactor left nanoswitch without a way to sink its logs as it doesn\u0027t have\na gRPC endpoint to read them from and logtree doesn\u0027t output to any stdout/stderr by default.\nThis adds a simple sink similar to what\u0027s currently used the node code.\n\nTest Plan: Manually tested\n\nX-Origin-Diff: phab/D672\nGitOrigin-RevId: 8698f9f920f36adf11fa1ef34a47723514eaa665\n" } ], "next": "662b5b3119b0798980b887d1ef9fa1b5632aa7fb" }