Updated live sync script

Discuss your pilot or production implementation with other Zimbra admins or our engineers.
umarzuki
Advanced member
Advanced member
Posts: 97
Joined: Sat Sep 13, 2014 12:46 am

Updated live sync script

Post by umarzuki »

How about deleting .ssh folder on both sides and re run that ssh key steps?
Sent from my Lenovo S920_ROW using Tapatalk
picalo
Posts: 5
Joined: Sat Sep 13, 2014 3:28 am

Updated live sync script

Post by picalo »

[quote user="twokeys"]This is super strange, really. But I feel like it's an Ubuntu thing for some reason. Anyways let's figure this out. Try this
You're on a development environment so you don't need to worry so much about security (I would assume..) so you don't need to worry too much about restricting the SSH session to that one "/opt/zimbra/live_sync/sync_commands" command.
On the primary server, do a cat /opt/zimbra/.ssh/live_sync.pub copy the contents, and paste it into authorized_keys on secondary server.
Now to test it, from the primary server as Zimbra, run ssh -i /opt/zimbra/.ssh/live_sync zimbra@ip_of_standby
If that doesn't work, I honestly have no more ideas because this is just standard SSH keys that you're having problems with.[/QUOTE]
Okay, this is what I will to as soon as I have gotten stuck again (restarting from scratch). I just feel like I am screwing something up, the process is clearly laid out and no one else seems to be having this issue. I will report back today with an update...again thank you very much for that assistance.
picalo
Posts: 5
Joined: Sat Sep 13, 2014 3:28 am

Updated live sync script

Post by picalo »

[quote user="umarzuki"]How about deleting .ssh folder on both sides and re run that ssh key steps?
Sent from my Lenovo S920_ROW using Tapatalk[/QUOTE]
This is something I will try as soon as I get stuck at that spot again (re-doing the entire set up). These are all great Ideas one of them has to fix this bugger.
umarzuki
Advanced member
Advanced member
Posts: 97
Joined: Sat Sep 13, 2014 12:46 am

Updated live sync script

Post by umarzuki »

Have anyone experienced redologs replaying more than one hour after more than 3 days outage on backup server?
Sent from my Lenovo S920_ROW using Tapatalk
16428simonb
Posts: 45
Joined: Fri Sep 12, 2014 9:55 pm

Updated live sync script

Post by 16428simonb »

Three days of redo logs will take a long time to play back on a busy server. Over one hour is not surprising. Try tailing /opt/zimbra/mailbox.log to see if the replay is actually happening.
umarzuki
Advanced member
Advanced member
Posts: 97
Joined: Sat Sep 13, 2014 12:46 am

Updated live sync script

Post by umarzuki »

I found a few ERROR "unable to execute redop" zimbra/log/mailbox.log
When I diverted zmplayredo to a file instead of /dev/null on live_sync script, I found that zmplayredo stated mailboxd is not running but it complained it is running when I started mailboxd with zmmailboxdctl start
16428simonb
Posts: 45
Joined: Fri Sep 12, 2014 9:55 pm

Updated live sync script

Post by 16428simonb »

Hmmm. Maybe there are processes running with an older PID to that recorded in the PID files. This happens to me sometimes with LDAP which fails to start because a previous LDAP process that Zimbra doesn't know about is already binding to the address/port. Issue a zmcontrol stop and then ps aux looking for any zimbra processes still running, especially java processes such as the one running mailboxd. If necessary zap the dangling Zimbra processes that shouldn't be there after a zmcontrol stop.
umarzuki
Advanced member
Advanced member
Posts: 97
Joined: Sat Sep 13, 2014 12:46 am

Updated live sync script

Post by umarzuki »

I also had tried
1- zmcontrol stop
2- made sure all zimbra related services stopped
3- removed zmmailboxd_manager.pid
4- zmsshkeygen & zmupdateauthkeys on all servers
but I'm still having mailboxd not running from zmplayredo
any more idea?
this is before my last resort to resync from live server which about 2 TB data (zimbra dir & HSM)
SevenQri
Posts: 5
Joined: Sat Sep 13, 2014 3:11 am

Updated live sync script

Post by SevenQri »

hi guys
i have been implemented this great Live_sync for several years ...

last month when i move to another room the zimbra server along with live_sync server i just realized that there was a problem ...

the live_sync could not be started ... error timestamp check failed

after googling i found the certificate was expired so i create new 10 years certificate

then got an error again ... invalid credential

googling again i found ldap password was no match so i matched it to live server
then i can start live_sync again ...

but ...

this morning i found no live_sync.log
then i tried to started live_sync again and found this log
[zimbra@mail live_sync]$ tail -f log/live_sync.log

***************************************

Mon Aug 4 09:05:25 WIT 2014 :Starting live_syncd

Mon Aug 4 09:05:25 WIT 2014 :Incremental backups enabled : false

Mon Aug 4 09:05:25 WIT 2014 :Convertd enabled : false

Mon Aug 4 09:05:25 WIT 2014 :Starting ldap live sync process

Mon Aug 4 09:05:25 WIT 2014 :Starting redo log live sync process

Mon Aug 4 09:05:25 WIT 2014 :Mon Aug 4 09:05:25 WIT 2014 :Ending redo log live sync process

Ending ldap live sync process


and after a few minutes the log file disappear
any clue ?
thanks
SevenQri
Posts: 5
Joined: Sat Sep 13, 2014 3:11 am

Updated live sync script

Post by SevenQri »

Solved with delete all folder in /opt/zimbra/live_sync
:)

thanks
Post Reply