Re: higher modseq on replica

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, 29 Mar 2021, at 4:51 PM, mtwomey via Info wrote:
I'm in the process of migrating a quite old Cyrus server from v2.4 -> v3. Right now, I'm on early steps and I've setup a 3.0 replica and I've been able to sync accounts without issue.

As part of my testing, I've browsed a few mailboxes with pine on the replica. However, this cause an update and now when I sync those users from the master server, I get a "higher modseq on replica" error and it bails out.

This is expected.  Clients should not connect to a replica, because they will change its state.  The replication (especially back in 2.4) expects that changes are only made to the master, and if the replica is not in the expected state the replication will, quite correctly, bail out rather than risk breaking something further.

I understand that you're testing 3.0 and need to see how it behaves when examined by a client, and that's fine; it just means that you should wipe these users and re-replicate them once you're done testing, before you proceed for real.

I think I can just sync_reset these users on the replica and re-sync them. However, I'm curious if I can "force" this to update (overwrite?) despite the change on the replica so that I don't need to wipe and redo these users?

In newer Cyrus versions there is some smarts for resolving "split brain" situations (i.e. where both sides of the pair have local changes), but I don't know much about it.  But I don't think it's in 2.4, so it won't help you here.

Cheers,

ellie

[Index of Archives]     [Cyrus SASL]     [Squirrel Mail]     [Asterisk PBX]     [Video For Linux]     [Photo]     [Yosemite News]     [gtk]     [KDE]     [Gimp on Windows]     [Steve's Art]

  Powered by Linux