Ext4 partition damaged with zimbra

[marciogentil]

Good Morning. Gentlemen, I ask a help to be able to identify why Linux Centos partitions are damaged. I realized that this happens in VM's that have database such as MySQL (zimbra). It does not happen VM's Windows. I use the CentOS version 6.7 64-bit and vSphere ESXi 6.0.0, 2494585. Minhaversão and Zimbra 8.6.0 path 7
My host is an HP DL 360 Gen9 and the VMDK is the host of the disc itself. Below, let an example of the message:
EXT4-fs error (device dm-0): ext4_lookup: deleted inode referenced: 7078651
EXT4-fs error (device dm-0): ext4_lookup: deleted inode referenced: 7078651
EXT4-fs error (device dm-0): ext4_lookup: deleted inode referenced: 7078651
EXT4-fs error (device dm-0): ext4_lookup: deleted inode referenced: 7078651
EXT4-fs error (device dm-0): ext4_lookup: deleted inode referenced: 7078651

[BradC]

That has absolutely nothing to do with ZImbra. You have a data consistency issue with your VM. You need to talk to the vendor of the VM OS and potentially the host OS.

I'd boot the VM from a live-cd and do a thorough check of the filesystem, then start searching the logs for any indication of issues, but this is strictly an issue in your infrastructure and not caused by Zimbra.

[marciogentil]

Brad, thanks for the reply. I also believe that Zimbra is not the problem. Just do not know what to search or analyze to find out what is casuando this problem of corrupt partition. I'll check with HP. Thank you

[tonster]

Also, it's just incorrect to say this only affects Linux. Yes, only Linux has ext4 but windows filesystems become corrupt too. Generally happens if the system is not shut down properly or the storage the vmdk is on its suddenly detached before the os is shut down.

Sent from my SM-G925T using Tapatalk

[marciogentil]

Tonster, thanks for the reply. I understood what he wrote, but VM Windows (2008 and 2012) are in the same HD that Linux VM. Linux is corrupt and windows is not corrupted.

I noticed the danificda partition, but zimbra database is damaged. I can not even remove an account in zimbra. In mysql_erro.log file, you are showing the following message:

InnoDB: http://dev.mysql.com/doc/refman/5.6/en/ ... overy.html
InnoDB: for more information.
160829 9:41:12 [ERROR] Got error 180 when reading table './mboxgroup38/open_conversation'
160829 9:41:12 [ERROR] Got error 180 when reading table './mboxgroup38/open_conversation'
2016-08-29 09:41:19 7f026e379700 InnoDB: Error: page 1414 log sequence number 10434589683
InnoDB: is in the future! Current system log sequence number 7441913700.
InnoDB: Your database may be corrupt or you may have copied the InnoDB
InnoDB: tablespace but not the InnoDB log files. See
InnoDB: http://dev.mysql.com/doc/refman/5.6/en/ ... overy.html
InnoDB: for more information.

Thanks

[andreaswolske]

I assume a hardware issue. What backend storage does your ESXi use? Local disk, iSCSI or NFS or even FC?
Do you have VMware Tools installed inside your Linux VM running Zimbra?
Cloud you please post /etc/fstab of your Zimbra VM?

Please boot your Zimbra VM with https://www.system-rescue-cd.org/System ... d_Homepage and check / repair the FS.
The innodb error does not necessarily point to a corrupt FS.
If the FS is O.K. add

Code: Select all

innodb_force_recovery = 6

to

Code: Select all

/opt/zimbra/conf/my.cnf

and check if InnoDB is able to repair the Logs after reboot.
After this has been done you can use zmblobchk to clean up orphaned files and DB entries to avoid error messages.
BR
Andreas