HD inode issues - now no Zim, install on top same issue

[BarefootPanda]

So I had my LVM freak out when I changed out a dying HD. I had a lot of inode issues, though fsck was able to repair them. The rest of the system appears in good shape, but the issues found were in the zimbra dir - sorry, didn't document issues in single user mode.

Anyways, I attempted to re-install over the top of my current install hoping that any damaged files would be restored - it didn't work. I'm in the same spot.

Running CentOS 5.4....Release 6.0.2_GA_1912.RHEL5_64_20091020161509 CentOS5_64 FOSS edition.

When I attempt zmcontrol stop I get...
Stopping stats...Terminated

and then a prompt.

I've tried zmfixperms, updating the system, reboots, etc. Where else should I start? Is there other info that would be helpful?

There are only a few test users on this box, but I'd like to attempt to repair the system if for nothing other than a good exercise.

[y@w]

I would call Dusty. Specifically in the middle of the night.


I don't know if this will be helpful, but I did find this: [SOLVED] zmcontrol not working

Anything useful in /var/log/zimbra.log or /opt/zimbra/log/mailbox.log?

[BarefootPanda]

Yeah Wyatt that's what he wants!! LOL I got the LVMs rebuilt, drives swapped and system running with very little downtime w/o him so I'm rather pleased.

I did the same stuff that guy did in your link yesterday afternoon - at this point I've tried re-installing about 3 or 4 times with no luck.

/var/log/zimbra.log mostly has stuff like...

Nov 12 07:32:08 mail zmmailboxdmgr[23780]: status OK
Nov 12 07:32:18 mail clamd[19555]: SelfCheck: Database status OK. 
Nov 12 07:34:08 mail zmmailboxdmgr[24254]: status requested
Nov 12 07:34:08 mail zmmailboxdmgr[24254]: status OK
Nov 12 07:36:08 mail zmmailboxdmgr[24728]: status requested
Nov 12 07:36:08 mail zmmailboxdmgr[24728]: status OK
Nov 12 07:38:08 mail zmmailboxdmgr[25201]: status requested
Nov 12 07:38:08 mail zmmailboxdmgr[25201]: status OK
Nov 12 07:40:01 mail zimbramon[25333]: 25333:info: 2009-11-12 07:40:01, QUEUE: 21 5 
Nov 12 07:40:10 mail zmmailboxdmgr[25905]: status requested
Nov 12 07:40:10 mail zmmailboxdmgr[25905]: status OK

I don't have a /opt/zimbra/log/mailbox.log - I only have things like....

# cat /opt/zimbra/log/mailbox.log.2009-11-
mailbox.log.2009-11-03  mailbox.log.2009-11-05  mailbox.log.2009-11-07  mailbox.log.2009-11-09  
mailbox.log.2009-11-04  mailbox.log.2009-11-06  mailbox.log.2009-11-08  mailbox.log.2009-11-10

For reference - when I run zmcontrol status after re-re-reinstallation I get this...

Host mail.domain.com
	antispam                Stopped
		zmmtaconfigctl is not running
		zmamavisdctl is not running
	antivirus               Stopped
		zmmtaconfigctl is not running
		zmamavisdctl is not running
	ldap                    Running
	logger                  Running
	mailbox                 Running
	mta                     Running
	snmp                    Running
	spell                   Stopped
		zmapachectl is not running
	stats                   Running

[BarefootPanda]

Really? No one has a quick, disaster recovery process for the GA FOSS release?

If I install and don't choose to upgrade will it destroy my user data? I'd assume so. Are there folder/files I can move out of the zim install, clean install and then move files back?

[BarefootPanda]

So, I ended up killing all zimbra processes since zmcontrol stop doesn't work.

I created a tar of my /opt/zimbra/ dir and then did a clean install. It worked beautifully.

I tried to stop the server, restore the tar, then run an upgrade install and I'm back to square one.

At any rate, it seems it's my /opt/zimbra dir and not the system.

Can I do a clean install and then restore my domain/user data somehow?

[ArcaneMagus]

I'm sure that I am forgetting at least something... but try these directories and see if it works? (all of these are in /opt/zimbra)
./data
./conf
./db
./logger
./store
./redolog
./ssl
./zimlets-deployed


Copy them over, upgrade install.. see what happens?

Also try creating an empty /opt/zimbra/log/mailbox.log with 640 zimbra:zimbra permissions, or running logrotate to make sure it creates the file.

[BarefootPanda]

Originally Posted by ArcaneMagus

I'm sure that I am forgetting at least something... but try these directories and see if it works? (all of these are in /opt/zimbra)
./data
./conf
./db
./logger
./store
./redolog
./ssl
./zimlets-deployed


Copy them over, upgrade install.. see what happens?

Also try creating an empty /opt/zimbra/log/mailbox.log with 640 zimbra:zimbra permissions, or running logrotate to make sure it creates the file.

Got me MUCH closer!

Did a clean install, stopped with zmcontrol stop, rsync -a ./yourlist /opt/zimbra/, then upgrade install and it seemed to behave differently than my previous upgrade installs. I stopped with zmcontrol and ran zmfixperms then restarted with zmcontrol start....

Host mail.domain.com
	Starting ldap...Done.
	Starting logger...Done.
	Starting mailbox...Done.
	Starting antispam...Done.
	Starting antivirus...Failed.
Starting zmmtaconfig...zmmtaconfig is already running.
Starting amavisd.../opt/zimbra/bin/zmamavisdctl: line 111: /opt/zimbra/amavisd/sbin/amavisd: No such file or directory
failed.
Starting freshclam...done.
Starting clamd...failed.


	Starting snmp...Done.
	Starting spell...Done.
	Starting mta...Password: 
Password: 
Password: 
Password: 
Failed.
Starting zmmtaconfig...zmmtaconfig is already running.
/opt/zimbra/bin/postfix: line 56: /opt/zimbra/postfix/sbin/postmap: No such file or directory
Sorry, try again.
Sorry, try again.
sudo: 2 incorrect password attempts
postfix failed to start
Starting saslauthd...already running.


	Starting stats...Done.

[ArcaneMagus]

Well I don't know if I would exactly call that closer, but it's something at least lol...

Hmm... any idea why it's talking about missing files? Those should be there from the install...

Also looks like there is some problem with the sudoers file.

What kind of stuff are you seeing in /var/log/zimbra.log or /opt/zimbra/log/mailbox.log (if that exists...)?

[BarefootPanda]

I did notice when I just attempted a re-install after doing a stop and then killing all zimbra processes as well that when it is "restoring from config..." it hits a snag on that /opt/zimbra/postfix/sbin/ dir. That directory doesn't exist...though it does in my backup. Would it be wise to rsync the postfix directory as well?

[ArcaneMagus]

Hmmm.... the initial install should have created it considering that is where all of the postfix binaries are stored, and without it the system obviously cant start.

Yea try copying it from the backup I guess...