[SOLVED] Zimbra on DRBD

[c_kalbhor]

We have installed Release 5.0.16_GA_2921.UBUNTU8_64 UBUNTU8_64
about a month back.

It was working fine till yesterday afternoon. We are able to receive mails.
But when we try to send emails we get error.

I checked /var/log/zimbra.log file it gave following message :
Sep 18 16:58:03 webmail zmmailboxdmgr[20879]: status requested
Sep 18 16:58:03 webmail zmmailboxdmgr[20879]: status OK
Sep 18 16:58:03 webmail zimbramon[20716]: 20716:info: 2009-09-18 16:58:01, STATUS: webmail.chpl.in: antispam: Running
Sep 18 16:58:03 webmail zimbramon[20716]: 20716:info: 2009-09-18 16:58:01, STATUS: webmail.chpl.in: antivirus: Running
Sep 18 16:58:03 webmail zimbramon[20716]: 20716:info: 2009-09-18 16:58:01, STATUS: webmail.chpl.in: ldap: Running
Sep 18 16:58:03 webmail zimbramon[20716]: 20716:info: 2009-09-18 16:58:01, STATUS: webmail.chpl.in: logger: Running
Sep 18 16:58:03 webmail zimbramon[20716]: 20716:info: 2009-09-18 16:58:01, STATUS: webmail.chpl.in: mailbox: Running
Sep 18 16:58:03 webmail zimbramon[20716]: 20716:info: 2009-09-18 16:58:01, STATUS: webmail.chpl.in: mta: Stopped
Sep 18 16:58:03 webmail zimbramon[20716]: 20716:info: 2009-09-18 16:58:01, STATUS: webmail.chpl.in: snmp: Running
Sep 18 16:58:03 webmail zimbramon[20716]: 20716:info: 2009-09-18 16:58:01, STATUS: webmail.chpl.in: spell: Running
Sep 18 16:58:03 webmail zimbramon[20716]: 20716:info: 2009-09-18 16:58:01, STATUS: webmail.chpl.in: stats: Running
Sep 18 16:58:30 webmail postfix/postqueue[21058]: fatal: Queue report unavailable - mail system is down

I tried to get the satus on zimbra control prompt

zimbra@webmail:~$ zmcontrol status
Host webmail.chpl.in
antispam Running
antivirus Running
ldap Running
logger Running
mailbox Running
mta Stopped
postfix is not running
snmp Running
spell Running
stats Running

Thus it gives mta Stopped and postfix is not running
Further zimbra is also not opening in browser using http://localhost/zimbra
Yes but it opens with https://localhost/zimbra

Kindly help - Chandrashekhar

[hizar2000]

Originally Posted by sems

Heartbeat does in R1 mode does not check the status of each resource, whether they are up or down. So, you'll need to implement something, like MON package to monitor zimbra status and if needed stop heartbeat accordingly.

Hizar,

In your setup, I haven't seen anything related to split-brain on DRBD. That has been my major issue (I have been running an active/passive cluster since March, but it always comes back as a problem). How do you deal with it on yours? I have been searching on the net for possible ways to prevent split-brain to happen in the first place. I found some articles suggesting to increase heartbeat's drbddisk script's retry from 6 to 20 and increasing timeout values (see Otherworldly Gate: STONITH with DRBD and Heartbeat), I haven't tried it yet but I am open to suggestions.

Yes this is an issue with my setup. I've never really understood how this whole Stonith thing is supposed to work and when I've tried it, it hasnt quite given me the results I was expecting. Other than a Stonith device, it seems hard to get stonith to work properly. (If the network between the two hosts is down, then how is one host supposed to kill the other?)

The other thing I tried that got me mixed results is setting up 'Shoot myself in the Head' which as it sounds is when the other host is unreachable, kill yourself.

This has the effect that both servers recover after a minimal downtime, but its hardly High Availability.

The other problem I had was sometimes it refused to work because something was holding the DRBD partition open and would not die.

[lassaad]

empty the qeue

[Himanshu]

Originally Posted by papaya74

Hi folks,

on my Debian-Servers i got ZIMBRA successfull working with heartbeat on a DRBD netraid !

Short description (how i did it): FOR TESTING ONLY !!!!
==============================================

I ASSUME, that you have a proper running installation of heartbeat & DRBD !!

1. Make 2 (!) installations of Zimbra with exactly the same parameters, on 2 different server, of course.

2. Define one server as Master (for yourself and later for heartbeat), and configure this one to your ZIMBRA-requirements.

3. Test your Master-Zimbra-Server if everything is fine, up & running

4. Remove on BOTH NODES ALL Zimbra auto-start-scripts, like

rm /etc/rc2.d/S99zimbra
rm /etc/rc3.d/S99zimbra
rm /etc/rc4.d/S99zimbra
rm /etc/rc5.d/S99zimbra

5. STOP ALL Zimbra services:

/etc/init.d/zimbra stop

6. MOVE /opt to somewhere else, e.g /opt_save, and create /opt again:

# mv /opt /opt_save
# mkdir /opt

7. Get the DRBD working on /opt as a netraid. I don't know if it will work with symlinks, so i decided to mount /opt directly with DRBD. Don't forget /etc/fstab

8. Now you have to reboot BOTH nodes, to be sure that you have a proper DRBD netraid running on /opt !!!
Check also if heartbeat is running as well and mounts /opt properly if your Master is dead !!

9. If everythings is working fine, then it's time to copy Zimbra back to production state:

mv /opt_save/zimbra /opt/

10. Edit /etc/ha.d/haresources on BOTH NODES and ADD 'zimbra' at the end of the line where you a controlling, what heartbeat should start & stop.

# nano /etc/ha.d/haresources

zimbra-01 drbddisk::netraid1 Filesystem::/dev/drbd0::/opt::ext3 YOUR_MASTER_IP zimbra

11. REBOOT both Nodes, and Zimbra should now start on the master. If the master dies, the 2nd node will takeover completly

BEWARE:
==> THIS WAS DONE JUST FOR TESTING PURPOSES !!!
==> EXACTLY ANALYSIS of Zimbra's behaviour is PENDING !!!

I hope it'll work in a production enviroment, i will test it this week with nearly 60GB email, so we will see....

/harry

Can u let me know if my Understanding is Correct on This..?

Server - 1 IP-1
Server - 2 IP-2
Virtual_Server IP-0

Direct Ethernet CrossOver Cable Connected between Server-1(IP-3)
& Server-2(IP-4) With No Switches .(For Replication & Hear Beat)

In Above Case Where IP-0,IP3 , IP4 Should be reflected.

[Himanshu]

Ok. I Got an article at

Howto: Highly available Zimbra cluster using Heartbeat and DRBD | mig5.net

Got it how this stuff would work,

[Samp]

And what news today with Zimbra and DRBD ?

Did you use it with LVM partitions ?

Thanks for the report.

Sam.

[Himanshu]

Yes.But The Mouting Would be Done as /dev/drbd .If you Stop drbd service & mount as LVM patrition you can access the data

[Samp]

Originally Posted by Himanshu

Yes.But The Mouting Would be Done as /dev/drbd .If you Stop drbd service & mount as LVM patrition you can access the data

Thanks for the advice.

Sam.

[Chadsel Chen]

DRBD with Zimbra 6 upgrade

Zimbra keeps almost everything in /opt/zimbra and very few system changes outside /opt/zimbra folder

What I did is I install a fresh 5.0.18 on my secondary DRBD site and upgrade to 6.0.4. After that when DRBD switch to secondary site, Zimbra start without problems.

Chadsel

[k_k]

hi guys,


this post is really interesting n useful..i have some queries :

1 ) how much data DRBD solution can support ??? any maximum limit ??
2 ) does this DRBD solution support storage appliance ? means if /opt/zimbra is mounted on storage partition then will it support ?

please help me to understand.