libre #smartos

24 Sep 2022

SmartOS -- A Modern, Open Source, Datacenter Operating System | smartos.org | github.com/TritonDataCenter/smartos-live | github.com/TritonDataCenter/triton | Security -- TritonDataCenter.com/security-issues | Channel logs - log.omnios.org/smartos

• 00:57
  tozhu
  does the zfs support ZSTD for smartos right now?
• 01:00
  jesse_
  uh... yes, but how?
• 07:40
  jfqd1
  jperkin: Got "pkgin: failed to commit transaction" while updating packages in a global zone. Any idea how to fix it?
• 10:07
  sjorge
  tozhu: I don't thing we have zstd yet, but not 100% sure
• 11:27
  tozhu
  sjorge: thank you
• 12:43
  jperkin
  jfqd1: what version?
• 12:57
  jfqd1
  jperkin: pkgin 22.9.0
• 12:59
  jfqd1
  repo is: ttps://pkgsrc.smartos.org/packages/SmartOS/trunk/tools/All
• 13:02
  jperkin
  jfqd1: anything useful in /opt/tools/var/db/pkgin/sql.log?
• 13:08
  jfqd1
  jperkin: SQL error at 2022-09-17 06:23:13 UTC
• 13:08
  jfqd1
  errmsg: near "??": syntax error in "?????????????;{"
• 13:08
  jfqd1
  query: ?????????????;{
• 13:08
  jfqd1
  Not that useful, right? :)
• 13:10
  jfqd1
  Should I remove the tools folder and start from scratch?
• 13:15
  jperkin
  nothing newer than that? I fixed issues in 22.9 where there were some corrupt queries
• 13:16
  jperkin
  so would expect to see them for previous runs but not newer, and I'd expect to see an error for the failed commit
• 13:16
  jperkin
  the transaction stuff is new in 22.9 so really interested to figure out why it's failing, haven't seen that yet in testing
• 14:23
  jfqd1
  jperkin: No, nothing newer in the log. The older parts of the log are here: gist.github.com/jfqd/b25dbd602a55e996a2d60095fd952db2
• 14:23
  jfqd1
  There error is only on this machine.
• 14:23
  jfqd1
  All others are fine.
• 14:57
  jperkin
  ok that's really interesting, still seeing corrupt queries but also transaction errors earlier, I'll have to see if I can reproduce
• 15:02
  jfqd1
  jperkin: Is there a better solution that removing the tools folder and start from scratch?
• 15:03
  jperkin
  you should be able to just move pkgin.db out of the way, if it works after that then I'd like a copy of the old one to see if there's anything in there causing problems
• 15:11
  jfqd1
  jperkin: moved pkgin.db and it worked again. Thanks a lot for your support!
• 15:11
  jfqd1
  Download of the db-files are here: nc2.qutic.com/s/JegmR3Qa6ZHJnoT
• 15:13
  jperkin
  nice, thanks
• 21:54
  jesse_
  trying bahamat's ubuntu 22.04 lx image, I get errors about
• 21:54
  jesse_
  $UUID/root/checkpoints: No such file or directory
• 21:54
  jesse_
  $UUID/root/etc/init/container-detect.override: cannot create [No such file or directory]
• 21:54
  jesse_
  $UUID/root/etc/init/mountall.override: cannot create [No such file or directory]
• 22:05
  bahamat
  That's odd.
• 22:06
  bahamat
  But the 20.04 one works fine?
• 22:09
  jesse_
  last time I deployed it, yes
• 22:09
  jesse_
  those are from stream":"zoneadmd
• 22:10
  jesse_
  is that error part of the vmadm plumbing?
• 22:10
  jesse_
  (is that part of the shell script in platform that is run every time lx zone is deployed (or started?)
• 22:11
  jesse_
  the full error: /usr/lib/brand/lx/lx_boot[104]: /zones/6240ea0e-ab15-c834-fbe1-b6d4eba93b13/root/etc/init/container-detect.override: cannot create [No such file or directory]
• 22:12
  bahamat
  It's because that init directory doesn't exist
• 22:13
  jesse_
  I dislike that script, it will fail whenever distros change or you want to run something else
• 22:13
  jesse_
  and it's hard to change as its part of platform
• 22:15
  jesse_
  maybe it should be an option in the manifest and by default works like now
• 22:19
  jesse_
  ...these init scripts were the problem I was having last year doing images, now that I look at them
• 22:19
  bahamat
  Where exactly do you see those errors?
• 22:20
  jesse_
  /zones/$UUID/logs/platform.log
• 22:20
  bahamat
  Does it cause anything to fail?
• 22:20
  jesse_
  the zone deploy fails
• 22:21
  bahamat
  I don't think those errors are causing the deploy to fail.
• 22:21
  jesse_
  well, it keeps dumping those errors to the log for some minutes and then fails
• 22:23
  bahamat
  That error happens on the 20.04 image, and it deploys just fine.
• 22:24
  jesse_
  console log has bunch of "2022-09-24T21:46:16.241Z INFO no metadata for 'sdc:volumes"
• 22:28
  bahamat
  It shouldn't, unless you've assigned a volume.
• 22:29
  jesse_
  I don't think so...
• 22:33
  jesse_
  bahamat, nothing suspicious in the manifest, pastebin.com/7XgMsuJm
• 22:59
  bahamat
  What platform are you running, by the way?
• 23:05
  jesse_
  joyent_20220210T011229Z
• 23:06
  jesse_
  I can update it if you think it'll help
• 23:41
  bahamat
  OK, 20.04 works on that PI, but 22.04 doesn't.
• 23:42
  bahamat
  I believe you might be running into OS-8372
• 23:42
  jinni
  smartos.org/bugview/OS-8372
• 23:42
  jesse_
  does 22.04 work on a newer PI?
• 23:42
  bahamat
  Yeah, it works on latest.
• 23:43
  bahamat
  If I'm right about the bug number, then you need at least 20220407
• 23:43
  bahamat
  And I'll need to set that in the image JSON before publishing it.
• 23:44
  jesse_
  "lucky" that I came across that one=)
• 23:45
  bahamat
  Well, anyone would have.
• 23:45
  bahamat
  And that's why I asked you to test it before I just published it.