Re: Hazelcast Stability Issues

[Enes Akar]

Can you try with 2.3? 

We should have been fixed this.

You can build it from master node 

or you can wait next week for release.

On Fri, Aug 24, 2012 at 1:02 AM, Lior Solomon <liors...@gmail.com> wrote:

Hello,

We've been using Hazelcast for quite a while now, and there is an issue we didn't manage to get rid of so far.

After restarting tomcat on our gateway machines everything works fine until we start getting the following warnings:

23/8 11:32:51 | WARN | (? ?:?) - () | [_._._._]:5707 [mygroupname] Handler -> RedoLog{name=c:__hz_Locks, redoType=REDO_TARGET_WRONG, operation=CONCURRENT_MAP_BACKUP_LOCK, caller=Address[_._._._]:5710 / connected=true, redoCount=84, migrating=null

partition=Partition [28]{

0:Address[_._._._]:5709

1:Address[_._._._]:5709

2:Address[_._._._]:5707

3:Address[_._._._]:5708

4:Address[_._._._]:5710

5:Address[_._._._]:5707

6:Address[_._._._]:5710

}

}

When this warning starts the servers are still responding but after a while everything halts and we are unable to access our services.

Only a tomcat restart will fix it.

Note: we have about 6 WAR files on each gateway's tomcat and each of these WARs is running an HazelCast instance.

We are not using multicast each server is configured to know the other gateways by IP.

We use only the HazelCast lock mechanism (no IMaps, no Queues etc).

Please advise.

Thanks in advance

Lior

--
You received this message because you are subscribed to the Google Groups "Hazelcast" group.
To view this discussion on the web visit https://groups.google.com/d/msg/hazelcast/-/ciD6-e-vWzsJ.
To post to this group, send email to haze...@googlegroups.com.
To unsubscribe from this group, send email to hazelcast+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/hazelcast?hl=en.

[Lior Solomon]

Hi Enes,

Thanks for the prompt response.

Can you please elaborate as little about the nature of the bug?

Thanks again,

Lior

[Enes Akar]

When there was a back-up operation on a partition that was scheduled to migrate, its address can not be realized and the redo operation was performed.

In fact this is not a serious bug, as scheduled migration is completed; operation is redone.

Anyway, bug is fixed on 2.3.

To view this discussion on the web visit https://groups.google.com/d/msg/hazelcast/-/PSSNq9928xYJ.