Question about what is replicated with Cyrus replication and what isn't

[egoitz at sarenet.es]

Hi mates, 

I have been taking a lot at the code and trying to debug this... 

I have seen there seems not to be replication aparently for neither
squatter nor cyr_expire. For instance, ipurge seems to be replicated
properly but not this way the two commented commands. I could launch
them in the slave, there's no problem with that but then, you see in the
logs the following lines : 

Feb 19 09:44:21 testserv imap[49589]: higher xconvmodseq on replica
___________________________________ - 1489 < 1496
Feb 19 11:45:14 testserv imap[99424]: higher xconvmodseq on replica
___________________________________ - 179 < 184
Feb 19 11:45:39 testserv imap[99424]: higher xconvmodseq on replica
___________________________________ - 537 < 614
Feb 19 11:45:40 testserv imap[99424]: higher modseq on replica
________________________________________ - 619 < 621
Feb 19 11:45:40 testserv imap[99424]: higher modseq on replica
________________________________________ - 625 < 627 

By what I have been reading previously, this is basically saying you
have made changes in the slave, which you shouldn't... because is that
exactly a slave.... but then... how do you handle delete_delayed and
expunge_delayed in a master/slave env?. Perhaps you should have always
expunge and delete mode to immediate in a slave always??. An later
before setting it as master, activate them again?. 

That could answer to cyr_expire question... but what about squatter and
xapian databases tier nightly movement of data and so?. 

How should all this be handled in a master/slave env?. It seems not to
be replicated with the replication protocol... so then... if it's not
but later replication throws that messages?... what's the proper way to
handle Squatter, Cyr_expire, perhaps ipurge although I think it
shouldn't in this case because ipurrge gets properly replicated due to
having a mailbox_close() call which then goes to a
mailbox_unlock_index() which does a sync_log_mailbox(). Shouldn't
Squatter and cyr_expire launch sync commands to sync log?. I think this
is not a bug because else at Fastmail you probably should have noted
that... but then I don't really know how to handle this situations
(cyr_expire, squatter) in a master/slave server.... 

Bron, Robert, Ellie mates :) could you please tell us something about
how to do it properly?.. 

Thanks a lot mates!!!

El 2019-02-17 19:22, egoitz at sarenet.es escribió:

> Hi! 
> 
> Previously (in 2.3 and older versions), cyr_expire and ipurge actions for instance where not replicated to the slave. So, you needed to launch them in both, the master and the slave. My question is, are now replicated as mailbox replication commands?. What about commands like Squatter -F for removing in Cyrus 3 from the database non existing mails?. Or when you move data from search tiers in Xapian?. Are those actions replicated or should be launched in the master and the replica?. 
> 
> Cheers! 
> ----
> Cyrus Home Page: http://www.cyrusimap.org/
> List Archives/Info: http://lists.andrew.cmu.edu/pipermail/info-cyrus/
> To Unsubscribe:
> https://lists.andrew.cmu.edu/mailman/listinfo/info-cyrus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.andrew.cmu.edu/pipermail/info-cyrus/attachments/20190219/ba52699c/attachment-0001.html>