activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "mandar.wanpal" <>
Subject Re: Lease based locker fails as slave starts when master is still up
Date Tue, 18 Dec 2012 11:26:21 GMT
Thanks for response. 

I had asked our sys admin team to recheck NTP clock synch settings and get
those repaired. So they are ok now. 

A question here. we only need both master and slave server's clocks in synch
or DB server clock also shall be in synch with both of them. From where does
the TIME field in ACTIVEMQ_LOCK table takes value. from db server clock or
from master/slave clocks.

I will try and get the AWR report or any logs if available for that time
frame from our DBAs and will post the same here.

One observation now. I tried running similar setup on ActiveMQ 5.8.0
SNAPSHOT. I kept it running for more than 24 hrs and both nodes run in
master slave properly unless both nodes stopped after some network glich and
IO error accessing DB. Is there any way we can compell AMQ process to keep
running for at least few secs, if DB connectivity is gone. so that there
isnt any outage.

I am still checking ActiveMQ 5.8.0 SNAPSHOT, to see if the issue with master
slave is resolved in it, and if it is, do you have any idea by when we will
have release version of ActiveMQ 5.8.0 available. I dont think putting
SNAPSHOT in production is a good idea.

or is there any way this AMQ 5.7 can be patched to resolve master slave
issue with code from AMQ 5.8

Mandar Wanpal
View this message in context:
Sent from the ActiveMQ - User mailing list archive at

View raw message