[Csync2] Two questions (same changes on 2 hosts, replicating removals to slave)

[Lars Ellenberg]

On Wed, Sep 14, 2011 at 04:27:53PM +0400, Тарас Савчук wrote:
> Hi,
> 
> I'm using csync2 to replicate data between 2 masters (with older/younger conflicts detection) and 2 slaves.
> 
> 
> 1)      Problem with same changes in file on 2 hosts
> 
> In the csync2 docs I see (1.1.1 section):
> 
> <It is not considered as a conflict by Csync2 when
> the same change has been performed on two hosts
> (e.g. because it has already been synchronized with
> another tool).>
> 
> But when I have 2 identical files with same modification time on both masters with older/younger conflicts auto detection, I have same error on both sides:
> ERROR from peer web2: File is also marked dirty here!
> Do not auto-resolve conflict: Lost 'younger/older' test.
> 
> Is it correct behavior or bug?

My understanding of your scenario:
  somefile <<- under csync2 control, all in sync
  sleep 2;
  echo blafoo >> somefile
  rsync -av somefile other-node:
  csync2 -x somefile
   --> errors out, due to same mtime, even though the files are
   identical, and should thus simply be considered as "nothing to do".

If I understand the scenario correctly,
I would consider that behaviour a bug.

Any volunteers?

> 2)      Replicating removals to slave
> 
> My configuration is: csync2 on slaves are running as daemon and never run by cron (as I understand from docs, it's senseless).
> Quantity of files on slaves are growing as they are never deleted (files deleted on masters are never deleted on slaves). How can I solve this problem?

Deletions are supposed to be replicated.
Can you provide a minimal test scenario, that shows this behaviour?

Can you check if this behaviour is consistent
accross csync2 versions (1.34 vs. 2.0), or a regression?

> Thanks in advance and thanks for great software!
> 
> 
> --
> Savchuk Taras
> 


-- 
: Lars Ellenberg
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com