Split brain problem in Version 9.4.2

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Split brain problem in Version 9.4.2

look
I encounter split brain issue. Can help tell what the root reason?
image 1, log from standby jms
image 2, log from active jms

Reply | Threaded
Open this post in threaded view
|

Re: Split brain problem in Version 9.4.2

look
will the time gap between two servers cause this issue?
Reply | Threaded
Open this post in threaded view
|

Re: Split brain problem in Version 9.4.2

look
In reply to this post by look
there also a log as below,
#date time#/sys$time/INFORMATION/System time has changed (delta=28617), reordering timer task queue
Reply | Threaded
Open this post in threaded view
|

Re: Split brain problem in Version 9.4.2

IIT Software
Administrator
System time change is described here. Might be related to the disconnect as timers might be fired at a different time and you might get a heartbeat timeout.

This "Disconnect in Progress, try later" is thrown when the previous TCP connection (the rep channel) is not fully disconnected. You will see this messages until the disconnect is finished.

"Software caused connection abort" is related and usually thrown when the server socket at the other instance has no space in its accept buffer, that is, cannot accept connections at this time ("Disconnect is in progress").

So my guess is that it is related to the system time change which is a delta of about 28 seconds.

Look here how to solve a Split Brain.
Reply | Threaded
Open this post in threaded view
|

Re: Split brain problem in Version 9.4.2

DouglasJD
Will the time server gap cause the split brain issue? if yes, what is the maximum time gap allowed?