Re: [Dovecot] Question about "Storage keeps breaking" in logs
This dovecot box (RHEL 5.6 x64, dovecot 2.0.13, ext4) is a VM over vmware vsphere. The remote SAN disks are presented via iSCSI to the ESX hosts, and the ESX host show them as "local storage" /dev/sdX
We have several VM sharing that datastore, no I/O contention detected. I will try with imaptest but this a production enviroment, I will take a couple of weeks to get a "mainteneance window" for this.
Any other idea about this? Maybe this problem is load-related?
Regards
Maria
----- Original Message ----- From: Timo Sirainen Sent: 06/28/11 01:33 AM To: Maria Arrea Subject: Re: [Dovecot] Question about "Storage keeps breaking" in logs
On Wed, 2011-06-22 at 19:32 +0000, Maria Arrea wrote: > Jun 6 17:03:23 buzon dovecot: imap(juanan): Warning: mdbox /buzones/18/66/juanan/storage: rebuilding indexes Problem 1: It doesn't log the reason for this rebuild. I can find only one place in code why that could happen, fixed: http://hg.dovecot.org/dovecot-2.0/rev/028b953e7040 > Jun 6 17:03:26 buzon dovecot: imap(juanan): Error: mdbox /buzones/18/66/juanan/mailboxes/P7.Bielorrusia/cd-eslpod-enviado/dbox-Mails: Storage keeps breaking Problem 2: It apparently can't fix that error. But in my tests it gets fixed fine.. > Jun 6 17:03:27 buzon dovecot: imap(juanan): Error: Recent flags state corrupted for mailbox Deleted Messages Probably related to the index rebuild. So this is a local disk with ext4 filesystem, nothing special? It shouldn't be breaking.. How often do these happen? Can you reproduce these by running imaptest against a test account? http://imapwiki.org/ImapTest
On Tue, 2011-06-28 at 11:01 +0000, Maria Arrea wrote:
Any other idea about this? Maybe this problem is load-related?
All I can really say is that: This just shouldn't be happening, ever. If there's a way to reproduce it I would be interested in knowing how. I know there are other mdbox users who haven't had any corruption trouble for a long time..
participants (2)
-
Maria Arrea
-
Timo Sirainen