[Gluster-devel] Network Outage - Expected Behaviour?

[Gordan Bobic]

On Thu, 09 Jul 2009 17:29:40 +0200, Daniel Maher <dma+gluster at witbe.net>
wrote:
> Gordan Bobic wrote:
>> What is the expected behaviour of gluster client when servers disappear
>> (due to a short network outage) and re-appear? For example, say there
are
>> servers A and B, and client C. A goes away (e.g. pull the network
cable).
>> Timeout occurs, client continues using B. A returns without glusterfsd
>> being restarted or the client glusterfs being restarted.
>> 
>> Does the client periodically re-scan for presence of servers that
dropped
>> out? Or does the client have to be restarted to notice that a server has
>> returned?
> 
> As witnessed during my last round of testing some time ago under 
> 2.0.0rc1 (yes, i know), client C will _not_ notice that server A has 
> returned.
> 
> This behaviour may have changed since, however...

That's what I _thought_ I noticed about 2.0.2 the other day, but wasn't
sure. After disconnecting (via network) one of the nodes, it looked like it
wasn't getting updates any more afterwards. But fully rebooting it made it
work again. It's possible that I just didn't want for long enough for the
period re-scan to find it again, though. What is the rescan period? Is it
adjustable?

Can any devs confirm if/when a fix for this went into the release? And can
anyone else confirm or deny this still exists in any recent release? I'll
retest on my setup over the weekend, but multiple test cases would probably
help for filing a bugzilla report.

Gordan