The fix for the double free thing has been pushed quite a few hours ago.
If you’re on 13901 or higher, then you have the fix.
Obviously not helping with broken systems as those still need to restart the containers or unmount lxcfs from them.
https://github.com/lxc/lxd-pkg-snap/commits/latest-candidate shows the upcoming changes to the snap. Note that something being in there doesn’t mean it’s in stable yet, but it does mean that the next stable update will include it.
I’ve just finished off restarting my LXD hosts and will monitor the situation.
One thing to note is that my SWAP numbers are completely off inside my containers:
On host:
# free -m
total used free shared buff/cache available
Mem: 32167 8416 14805 65 8946 23356
Swap: 8191 35 8156
Inside containers:
# lxc exec container -- free -m
total used free shared buff/cache available
Mem: 953 368 108 0 477 585
Swap: 8796093021254 0 8796093021254
Also, unrelated to LXD, snap changes gives me an error whereas just yesterday it showed me some changes:
# snap changes
error: no changes found
# snap list
Name Version Rev Tracking Publisher Notes
core 16-2.43.3 8689 stable canonical✓ core
go 1.13.8 5364 stable mwhudson classic
lxd 3.22 13901 stable canonical✓ -
Is there anything special I need to do on my end to resolve this SWAP issue? Prior to recent updates, this issue appeared to be present on 18.04 only. But it is now plaguing both platforms.
Not sure what this would be about but I’ve improved the meminfo codepath now that should handle corner-cases better.
Heads up:
After just 6 days of uptime, the /proc mount issue has returned for containers on my 16.04 host. No issues identified yet for containers on my 18.04 host.
grep lxcfs /var/log/syslog please
Apologies for the omission.
$ sudo grep lxcfs /var/log/syslog
Mar 24 07:23:41 lxcserver lxd.daemon[2753]: Closed liblxcfs.so
Mar 24 07:23:41 lxcserver lxd.daemon[2753]: Running destructor lxcfs_exit
Mar 24 07:23:41 lxcserver lxd.daemon[2753]: Running constructor lxcfs_init to reload liblxcfs
Mar 26 06:49:04 lxcserver kernel: [519660.024260] lxcfs[32252]: segfault at 0 ip 00007f02baec6a32 sp 00007f02a1ffa6b0 error 6 in libc-2.23.so[7f02bae50000+1c0000]
Mar 26 17:38:44 lxcserver lxd.daemon[24169]: Running constructor lxcfs_init to reload liblxcfs
And snap version:
$ snap list lxd
Name Version Rev Tracking Publisher Notes
lxd 3.23 13987 stable canonical✓ -
Snap changes:
$ snap changes lxd
ID Status Spawn Ready Summary
233 Done yesterday at 17:38 AEDT yesterday at 17:38 AEDT Auto-refresh snaps "core18", "lxd"Ok, could you pastebin the entire crash entry that showed up around 06:49 in your syslog?
Interestingly, I see no other entry in the logs related to this crash - just that one line. Other than syslog, is there another log file that may have this info?
Log file kern.log also has just this one line.
Maybe journalctl -u snap.lxd.daemon -n 300 will show more?