MTA on Cluster HA Setup

[RAGHU_K]

Hi!,

I have 2 queries regarding the MTA configured on a Cluster running in HA mode:

1) What will happen to the Mails queued for processing in one member if there is a fail-over to the other member in the cluster? Is this queue will also be synced between the cluster members? If not then how to make sure the mails been delivered to the end-user?

2) How to address the scenario, when we face a hardware issue on the Active member in the cluster which is processing the mails? how to take out the mails which are in the queue?

Raghu

[TP_Master · Accepted solution]

Hi Raghu,
The MTA queues are not synced between members. That said, even if failed-over the MTA will keep processing the mails and deliver the remaining mails in its queue.

If there's a hardware issue (2) then you would need to manually copy the files from the MTA queues into the other member and re-inject them. I can help with instructions.
Thus far I haven't heard of any customer of ours that encountered this situation.

HTH

[Haoi_Yusuhriy]

please send the instructions

[RAGHU_K]

Hi! HTH,

Can you please send me the procedure for the second scenario, where one member is down & we need to inject the mails in the queue to the other working member?

Thanks,

Raghu

[RAGHU_K]

Hi! HTH,

Awaiting for your reply on the procedure.😊

I have tried my own method & on the Working Cluster Member, am able to see the mail queue of the problematic member. But unable to send the mails to the SMTP server manually. Getting an error "postqueue: fatal: Cannot flush mail queue - mail system is down"

Please find the below command output FYR:

======================================================================================================

[Expert@FW1-R8020:0]# /opt/postfix-fw2/postfix/usr/sbin/postqueue -c /opt/postfix-fw2/postfix/etc/postfix -p
postqueue: warning: /opt/postfix-fw2/postfix/etc/postfix/main.cf, line 9: overriding earlier entry: config_directory=/opt/postfix-fw2/postfix/etc/postfix
Mail queue is empty
[Expert@FW1-R8020:0]#
[Expert@FW1-R8020:0]# vi /opt/postfix-fw2/postfix/etc/postfix/main.cf
[Expert@FW1-R8020:0]#
[Expert@FW1-R8020:0]#
[Expert@FW1-R8020:0]# /opt/postfix-fw2/postfix/usr/sbin/postqueue -c /opt/postfix-fw2/postfix/etc/postfix -p
postqueue: warning: Mail system is down -- accessing queue directly
----Queue ID----- --Size-- ---Arrival Time---- --Sender/Recipient------
47t8qG6p6tz1Qr3X 24190620 Wed Jan 8 18:44:50 user100@rkmail.tk
(connect to 10.10.10.80[10.10.10.80]:25: No route to host)
user2@rkmail.tk

-- 23623 Kbytes in 1 Request.
[Expert@FW1-R8020:0]#
[Expert@FW1-R8020:0]#
[Expert@FW1-R8020:0]# /opt/postfix-fw2/postfix/usr/sbin/postqueue -c /opt/postfix-fw2/postfix/etc/postfix -f
postqueue: fatal: Cannot flush mail queue - mail system is down
[Expert@FW1-R8020:0]#

======================================================================================================

[Wolfgang]

RAGHU_K,

I suggest to open a TAC case to solve your problem as fast as possible.

This forum consist of a bunch of volunteers and some Check Point employees, there is no SLA here.

To your question:

You can simple copy all the files from the postfix queue directories (defferred, bounced etc.) to the other cluster node.

And then did an " /opt/postfix/usr/sbin/postqueue -c /opt/postfix/etc/postfix/ -f", which delivers all your messages.

Your shown message "postqueue: warning: Mail system is down " shows that the MTA/postfix is not running. You have to 

check why. Without this running you can't send any of your messages.

With "/opt/postfix/usr/sbin/postfix -c /opt/postfix/etc/postfix reload" you can try to restart postfix. If it not starting check

ATRG: Mail Transfer Agent (MTA)  for troubleshooting.

Your output shows only one message, which is not delivered. Maybee this message is really important for you.

Wolfgang