Commit Graph

1285 Commits

Author SHA1 Message Date
Miroslav Crnic 75dfd723c0 shuckle: fix ClearCdcInfoReq name 2024-09-17 10:05:46 +00:00
Miroslav Crnic b2ea95091a shuckle: support cdc replica moving across hosts 2024-09-16 17:31:47 +01:00
Miroslav Crnic c294cb8366 scratch: lifetime for scratch file 3 hours, heartbeating not needed 2024-09-16 15:21:28 +01:00
Miroslav Crnic 59fc480e85 shuckle: remove unused requests 2024-09-16 15:21:06 +01:00
Miroslav Crnic e41d01a7ce bufpool: add safety around double put and use after put 2024-09-16 11:07:33 +01:00
Nick Chapman d65ee658b0 Log hardware events to HDB via buildserver. (#338)
This updates eggsblocks to send hardware events to buildserver when a
serious disk error is encountered. Right now, all of the existing
alerting stays in place, but hopefully once we're confident that this is
working we can remove the alerts and rely on this automation instead.

There are two temporary alerts with the NCHAPMA: prefix to provide some
insight into whether this process is working.
2024-09-16 10:41:04 +01:00
Nick Chapman 0439db302e Directly write the build id
There's no point in writing a file that we have to parse later.
2024-09-13 17:47:08 +01:00
Nick Chapman 10c32ce63b Include build info in the eggs package 2024-09-13 12:19:30 +01:00
Miroslav Crnic 3adf331861 eggsblocks: checkBlock treat io.EOF from conversion as BAD_BLOCK_CRC 2024-09-12 14:57:21 +00:00
Nick Chapman dda8c33c1c Move files for deployment instead of copying 2024-09-12 15:42:17 +01:00
Nick Chapman cdc326225f Correct issue reading environment value 2024-09-12 15:29:24 +01:00
Nick Chapman 932aeda8b4 Correct issue with generating build numbers 2024-09-12 15:21:13 +01:00
Nick Chapman ee42e8429f Update build-eggsfs.yml 2024-09-12 15:08:09 +01:00
Miroslav Crnic 2b738e01c7 shard/cdc: location output in log 2024-09-12 14:06:38 +00:00
Nick Chapman db3e8da10a Rename build-kmod.yaml to .yml for consistency. 2024-09-12 15:05:39 +01:00
Nick Chapman 99f26770be Add environment option to build workflow. 2024-09-12 15:03:05 +01:00
Miroslav Crnic 1dc770f15f shard: format app name to be same as service name 2024-09-12 13:52:31 +00:00
Nick Chapman 37ff2131c4 Update build-eggsfs.yml
Correct glitch with generating package names
2024-09-12 14:51:01 +01:00
Miroslav Crnic 2dec9ec117 cdc: register location 2024-09-12 14:27:55 +01:00
Miroslav Crnic f8c4096dc9 shard: register location 2024-09-12 14:27:27 +01:00
Nick Chapman ccbeecfd25 Update build-eggsfs.yml 2024-09-12 14:21:43 +01:00
Nick Chapman f1fb391522 Update build-eggsfs.yml 2024-09-12 13:43:56 +01:00
Saulius Grusnys cbcd9f8162 handle edge case when restoring list order with padding pages (#340) 2024-09-12 09:03:48 +01:00
Nick Chapman c9792feaeb Create build-eggsfs.yml (#339) 2024-09-11 18:38:47 +01:00
Miroslav Crnic 4e11e1a00b shuckle: fix CDC_WITH_REPLICAS select 2024-09-11 16:30:17 +00:00
Miroslav Crnic 824972efcd shuckle: cleanup table conversion 2024-09-11 16:04:21 +00:00
Miroslav Crnic 8ac93a4c54 shuckle: add location for all services 2024-09-11 16:59:19 +01:00
Miroslav Crnic 90a35d67bb eggsshuckle: only alert on decommed services with files after 36h 2024-09-11 11:52:52 +00:00
Miroslav Crnic 73fc720ca1 eggsblocks: return io.EOF from conversion code 2024-09-11 08:54:08 +00:00
Nick Chapman 56b9dc192b Bump go version 1.18 --> 1.22. (#335)
The main change here is the updated go version in go.mod and the new
Docker build images which explicitly install go1.22 instead of relying
on the default 1.18 installation.
2024-09-10 17:05:03 +01:00
Miroslav Crnic 60bcf62f45 migrate: add option to only log which files would be migrated 2024-09-05 09:21:58 +00:00
Miroslav Crnic 066c32906e xmon: correctly reset time between reads 2024-09-05 08:31:47 +00:00
Miroslav Crnic f379afc168 scratchFile: log creation/destruction 2024-09-05 08:14:31 +00:00
Miroslav Crnic 450ba0f871 xmon: log error if we are delayed reading xmon connection 2024-09-04 14:18:47 +00:00
Miroslav Crnic d5522678ef eggstests: use old/new fetch block api at random to test both 2024-09-03 18:14:43 +01:00
Miroslav Crnic 0920c69099 shuckle: evenly distribute block services across shards 2024-09-03 17:44:32 +01:00
Saulius Grusnys 335659c0bf kmod: do not retry readahead requests 2024-09-03 12:44:55 +01:00
Miroslav Crnic 76695ec22b shard: check enough current block services 2024-09-02 16:49:31 +00:00
Saulius Grusnys 039fac6411 kmod: simplify restoring list after incomplete block fetch 2024-09-02 17:11:09 +01:00
Miroslav Crnic 35fe242fe3 eggsshard: check shuckle gives services from different fd-s 2024-09-02 16:05:59 +01:00
Saulius Grusnys e06b3463f9 retry logic in page reads (#330) 2024-09-02 11:15:32 +01:00
Miroslav Crnic 8171a3c7f0 eggsblocks: removal of without_crc will be done at end 2024-09-02 08:27:31 +00:00
Miroslav Crnic 400f93db18 eggsblocks: remove folder conversion 2024-08-30 15:30:42 +01:00
Miroslav Crnic 0c91d97ee7 cleanup: extract scratch file and make it more robust 2024-08-29 22:37:16 +01:00
Saulius Grusnys 0ac561b767 kmod: read inline files the old way, handle failed page allocation 2024-08-29 18:19:58 +01:00
Miroslav Crnic 802ecea6f4 kmod: prevent undeflow when initalizing block_services_last_changed_time 2024-08-29 16:58:19 +01:00
Miroslav Crnic a9759fec3f shuckle: cleanup 2024-08-29 14:47:16 +00:00
Miroslav Crnic 8d10036a00 shuckle: resolve flag propagation race 2024-08-29 14:51:42 +01:00
Miroslav Crnic e6e0662290 Revert "remove REDACTED from alerts/scripts temporarily"
This reverts commit 84d58291feb1b57a2af920fc53f7cc57c19c1d0c.
2024-08-29 13:00:10 +00:00
Miroslav Crnic 1edc459f1b remove REDACTED from alerts/scripts temporarily 2024-08-29 12:05:40 +00:00