[lxc-users] LXD 3.1: Raft: Failed to contact...

[Free Ekanayaka]

Hello,

first, you should use at least 3 nodes with a LXD cluster, otherwise if
just single node goes down (even for a reboot) you're cluster is
unavailable, because of lack of quorum.

Regarding "lxc ..." commands, there are improvements under work that
should improve performance quite a bit, likely to be released in LXD
3.3.

Finally, the raft error is very likely related to this issue:

https://github.com/lxc/lxd/issues/4548

for which a mitigation is present in LXD 3.2 (which just got released),
waiting for the larger fix in 3.3. That mitigation will also make your
"lxc ..." commands more predictable in timings (no fluctuation),
although for real improvements you'll have to wait 3.3, as said.

So please upgrade to LXD 3.2, and if you still hit those issues let us
know.

Regards,

Free

Vlad <vovan at vovan.nl> writes:

> Hello,
>
> I'm trying to setup 2 nodes LXD cluster and have some issues. Sometimes
> "lxc ..." command takes 5-15 seconds to complete, so I checked log's and
> I see every 10-15 minutes messages similar to the this one: "Raft:
> Failed to contact 2 in 1.500151s" following by "Raft: Failed to contact
> quorum of nodes, stepping down". The "lxc cluster list" command shows
> that both nodes are "fully operational". Another thing is the amount of
> TCP connections between two nodes:
> netstat -tnap  | grep lxd | wc -l
> 186
>
> Is this behavior expected or I'm missing something in configuration? I'd
> appreciate any thoughts about the issues above.
>
> Regards,
> Vlad.
> _______________________________________________
> lxc-users mailing list
> lxc-users at lists.linuxcontainers.org
> http://lists.linuxcontainers.org/listinfo/lxc-users