Hi, when upgrading lxd to 3.0.0 the lxc client can't connect to the local instance anymore: Error: Get http://unix.socket/1.0: dial unix /var/lib/lxd/unix.socket: connect: no such file or directory Accessing remote image servers seems to work though, as "lxc image list images:" works correctly. Maybe a bump of app-emulation/lxc is necessary, too? Reproducible: Always Steps to Reproduce: 1.sudo emerge --ask '=app-emulation/lxd-3.0.0' 2.sudo /etc/init.d/lxd start || sudo /etc/init.d/lxd restart 3.lxc ls
Created attachment 526328 [details] emerge --info lxc lxd
The missing socket means the daemon isn't running. You can get useful output by running the daemon in the foreground. In a root shell run: /usr/sbin/lxd --group lxd --verbose --debug I expect you'll see a clear error message before it exits, which should help you get unwedged. If not, you can post the output here and I'll try to help. I haven't reproduced this issue.
You're right, the daemon crashed. I didn't notice as OpenRC just prints the OK: % sudo /etc/init.d/lxd start lxd | * Starting lxd service ... [ ok ] % sudo /etc/init.d/lxd status * status: crashed It doesn't crash with the manual startup through your debug command: (I know of the cgroups errors, but they shouldn't be critical here as lxd2 has no problem with them. IMHO the sqlite errors sound weirder.) % sudo /usr/sbin/lxd --group lxd --verbose --debug INFO[04-03|19:12:36] LXD 3.0.0 is starting in normal mode path=/var/lib/lxd INFO[04-03|19:12:36] Kernel uid/gid map: INFO[04-03|19:12:36] - u 0 0 4294967295 INFO[04-03|19:12:36] - g 0 0 4294967295 INFO[04-03|19:12:36] Configured LXD uid/gid map: INFO[04-03|19:12:36] - u 0 1000000 65537 INFO[04-03|19:12:36] - g 0 1000000 65537 WARN[04-03|19:12:36] AppArmor support has been disabled because of lack of kernel support WARN[04-03|19:12:36] Couldn't find the CGroup blkio, I/O limits will be ignored. WARN[04-03|19:12:36] Couldn't find the CGroup CPU controller, CPU time limits will be ignored. WARN[04-03|19:12:36] Couldn't find the CGroup CPUacct controller, CPU accounting will not be available. WARN[04-03|19:12:36] Couldn't find the CGroup CPUset controller, CPU pinning will be ignored. WARN[04-03|19:12:36] Couldn't find the CGroup devices controller, device access control won't work. WARN[04-03|19:12:36] Couldn't find the CGroup memory controller, memory limits will be ignored. WARN[04-03|19:12:36] Couldn't find the CGroup network class controller, network limits will be ignored. WARN[04-03|19:12:36] Couldn't find the CGroup pids controller, process limits will be ignored. WARN[04-03|19:12:36] CGroup memory swap accounting is disabled, swap limits will be ignored. INFO[04-03|19:12:36] Initializing database gateway INFO[04-03|19:12:36] Start database node address= id=1 INFO[04-03|19:12:37] Raft: Initial configuration (index=1): [{Suffrage:Voter ID:1 Address:0}] INFO[04-03|19:12:37] Raft: Node at 0 [Leader] entering Leader state INFO[04-03|19:12:37] LXD isn't socket activated INFO[04-03|19:12:37] Starting /dev/lxd handler: INFO[04-03|19:12:37] - binding devlxd socket socket=/var/lib/lxd/devlxd/sock INFO[04-03|19:12:37] REST API daemon: INFO[04-03|19:12:37] - binding Unix socket socket=/var/lib/lxd/unix.socket INFO[04-03|19:12:37] - binding TCP socket socket=1.2.3.4:8443 DBUG[04-03|19:12:37] Initializing and checking storage pool "default". DBUG[04-03|19:12:37] Initializing a DIR driver. DBUG[04-03|19:12:37] Checking DIR storage pool "default". DBUG[04-03|19:12:37] Initializing a DIR driver. DBUG[04-03|19:12:37] Connecting to a remote simplestreams server DBUG[04-03|19:12:37] Connecting to a remote simplestreams server DBUG[04-03|19:12:37] Initialized inotify with file descriptor 14 INFO[04-03|19:12:37] Pruning expired images INFO[04-03|19:12:37] Done pruning expired images INFO[04-03|19:12:37] Updating instance types INFO[04-03|19:12:37] Expiring log files INFO[04-03|19:12:37] Updating images INFO[04-03|19:12:37] Done expiring log files DBUG[04-03|19:12:37] Database error: &errors.errorString{s:"sql: no rows in result set"} DBUG[04-03|19:12:37] Database error: &errors.errorString{s:"sql: no rows in result set"} DBUG[04-03|19:12:37] Database error: &errors.errorString{s:"sql: no rows in result set"} INFO[04-03|19:12:37] Done updating images INFO[04-03|19:12:39] Done updating instance types ^C INFO[04-03|19:13:23] Received 'interrupt signal', exiting. INFO[04-03|19:13:23] Stopping REST API handler: INFO[04-03|19:13:23] - closing socket socket=1.2.3.4:8443 INFO[04-03|19:13:23] - closing socket socket=/var/lib/lxd/unix.socket INFO[04-03|19:13:23] Stopping /dev/lxd handler INFO[04-03|19:13:23] - closing socket socket=/var/lib/lxd/devlxd/sock INFO[04-03|19:13:23] Closing the database INFO[04-03|19:13:23] Stop database gateway INFO[04-03|19:13:23] Stop raft instance INFO[04-03|19:13:23] Stopping REST API handler: INFO[04-03|19:13:23] Stopping /dev/lxd handler INFO[04-03|19:13:23] Stopping REST API handler: INFO[04-03|19:13:23] Stopping /dev/lxd handler INFO[04-03|19:13:23] Unmounting temporary filesystems INFO[04-03|19:13:23] Done unmounting temporary filesystems INFO[04-03|19:13:23] Saving simplestreams cache INFO[04-03|19:13:23] Saved simplestreams cache
Comparing your output to mine, I don't see any real errors there. I'm sure the sql "no rows in result set" messages can be ignored. My daemon, like yours, starts serving requests immediately after "Done updating instance types": ... INFO[04-03|18:33:16] Done updating instance types ^CINFO[04-03|20:20:52] Received 'interrupt signal', exiting. ... When running the daemon in the foreground, are you able to interact normally in another (non-root) terminal? e.g. 'lxc ls' etc. If so, then your issue is specific to the init script. Do you have anything non-standard in /etc/conf.d/lxd? Or, do you maybe have an old pidfile that needs to be deleted? /run/lxd.pid Finally, if you have lxcfs check if it the lxcfs daemon starts successfully. I just noticed that mine does not until I 'mkdir /var/lib/lxcfs'. I'll upgrade lxcfs this week and fix that issue at the same time.
It crashes with syslog logging enabled! "user % /usr/sbin/lxd --help" doesn't show --syslog in lxd-3.0.0. Did upstream maybe drop the support for it? I’m running app-admin/syslog-ng-3.14.1 with USE="caps", and AFAICS every local user is able to send messages through "logger 'lxd related logging test.'". (FTR: While trying to find the issue I bumped lxc and lxcfs each to 3.0.0 locally and both seem to run fine on my machine so far.)
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=aef61f3e7622e190f8833f0738661647c279a4f9 commit aef61f3e7622e190f8833f0738661647c279a4f9 Author: Erik Mackdanz <stasibear@gentoo.org> AuthorDate: 2018-04-08 20:04:56 +0000 Commit: Erik Mackdanz <stasibear@gentoo.org> CommitDate: 2018-04-08 20:04:56 +0000 app-emulation/lxd: Support latest confd options Closes: https://bugs.gentoo.org/652206 Package-Manager: Portage-2.3.26, Repoman-2.3.7 app-emulation/lxd/files/lxd.confd.1 | 24 ++++ app-emulation/lxd/lxd-3.0.0-r1.ebuild | 251 ++++++++++++++++++++++++++++++++++ app-emulation/lxd/metadata.xml | 4 +- 3 files changed, 277 insertions(+), 2 deletions(-)
Good catch. --syslog is no longer supported (and --print-goroutines-every was renamed). I've pushed -r1 with this fix.
