[SOLVED] Anyone else having trouble restoring mail accounts on 6.0.5?

[Jesster]

Hey Everyone,

I'm having zero success restoring any mailbox account on ALL our 6.0.5 mailbox servers. Is anyone else having issues? The backups are successful, and now we need to restore a few accounts. Anyways, I've already opened a Support Ticket, so we'll see if they uncover anything obvious (mailboxd restarts didn't fix it).

Java is throwing all kinds of tracebacks- heres a few:

Caused by: com.mysql.jdbc.exceptions.MySQLIntegrityConstraintViolationException: Cannot add or update a child row: a foreign key constraint fails (`mboxgroup29/mail_item`, CONSTRAINT `fk_mail_item_mailbox_id` FOREIGN KEY (`mailbox_id`) REFERENCES `zimbra`.`mailbox` (`id`))

and

Caused by: com.zimbra.common.service.ServiceException: system failure: writing new object of type 5
ExceptionId:LmtpServer-20:1267676645098:dc967d3c5dbedcba
Code:service.FAILURE

We've tried a random sample accounts, and all have failed:

[zimbra@zmbx02 ~]$ time zmrestore -a username@example.org -ca -pre restore_
Error occurred: Read timed out

real    5m3.001s
user    0m4.733s
sys     0m0.301s
[zimbra@zmbx02 ~]$

[Jesster]

Also experiencing this "Read timed out" on mailbox moves:

[zimbra@zimbra]$ time (zmmailboxmove -a username@example.org -t server2.example.org)
Error occurred: Read timed out

real    5m2.333s
user    0m4.189s
sys     0m0.232s

[LMStone]

Originally Posted by Jesster

Also experiencing this "Read timed out" on mailbox moves:

[zimbra@zimbra]$ time (zmmailboxmove -a username@example.org -t server2.example.org)
Error occurred: Read timed out

real    5m2.333s
user    0m4.189s
sys     0m0.232s

FWIW we are getting the same read time out errors on 6.0.5 larger mailbox moves, but the mailbox move completes OK in the background. We just refresh the Admin Console and wait until the mailbox status changes from "Maintenance" to "Active".

Haven't tried restoring a mailbox yet, but will do so and report back in the next day or so.

All the best,
Mark

[Jesster]

Thanks LMStone,

I noticed my tests are with accounts 800MB or greater. They are all successful regardless of the CLI error. I am working with support on this, and it sounds like they have a 5 minute Soap Timeout

zmlocalconfig httpclient_soaphttptransport_so_timeout

mine is set to 300000 which Support says is 5 minutes
to set the timeout to Infinite, just set it to 0. I will do some testing today.

--
Jessie

[ewilen]

Any update on this? I haven't tested but it'd be good to know if this is a common issue.

[Jesster]

Originally Posted by ewilen

Any update on this? I haven't tested but it'd be good to know if this is a common issue.

Hey Elliot,

See my previous post - we needed to change the timeout for the soap stuff to 0, the 6.0.x default is 5 minutes. I'm not sure why others aren't complaining. maybe they are restoring smaller mailboxes.

[ewilen]

Ah, so after changing the default, you got rid of the (cosmetic) error. Thanks.

[rusty]

zmlocalconfig httpclient_soaphttptransport_so_timeout

Running this command shows my timeout to be 300000 as well.

What is the command to actually change the timeout value?

Is it better to increase the value to 10 or 20 minutes? Seems like setting it to infinite defeats the purpose of a timeout.

Eric