backupd and sync_client IOERROR

Marco falon at
Wed Jul 8 08:08:15 EDT 2020

Hi Ellie,

On 08/07/2020 06:23, ellie timoney has written:
> Oh that's very curious.  It suggests that something about the mailbox contents is causing it to fail under -A but not under -u.  I was hoping it would just be the dot in the name, cause something like that should be fairly easy to reproduce and debug.  But it's sounding like it'll be more difficult!
> It might be worth running "mbexamine" over some of these mailboxes -- both some that work correctly, and the ones that are misbehaving -- and see if any interesting patterns emerge?  The man page is here:

The mailbox which fails is quite large (over 32000 msgs and over 1GB). 
mbexamine produces a very large output for all mails. I can see that it 
doesn't fail: it terminates without errors and with 0 exit status.

Even switches -c or -q terminates without errors. All quotas checks are 

>>> Does it fail in the same way replicating to a normal replica, instead of a backup server?
>> At this moment I don't have a normal replica server yet. I'll try...
> Thanks. I think this will be important for figuring out whether the problem is replication generally, or backupd specific.

I understand. As soon as I can, I'll setup a normal replica.

Thank you


More information about the Info-cyrus mailing list