[XI 3.0 SP21] Weird behaviour on FTP Sender adapter

[Former Member]

Hi All,

Lately I have been notticing some weird behaviour on our FTP sender channels. It seems that the channels are active on all server nodes of our J2EE stack at the same time.

For instance, we have an FTP sender channel that is active on a local adapter engine with 3 server nodes. Looking at the channel in the monitoring, I can see that the channel is active and polling on all 3 server nodes! Normal NFS channels do have normal behaviour however, polling only on one of the server nodes...

Any idea how this might be possible? We have implemented the clusterSyncMode = lock on all channels, but the behaviour persists... Stopping, starting, reactivating, rebuilding, nothing works...

Any help would be greatly appreciated!

Best regards,

Yordy

[Former Member · Accepted solution]

Sorry, but solved already. It seems that the "lock" value has some weird behaviour for clusterSyncMode.

If we leave the global value on Scheduler (in the adapter in the Visual Admin) and delete the parameter clusterSyncMode on the channel, it works as expected.

If we put the global value on lock, it only works if you put the channel value on scheduler, in no other configuration does it work...

So probably a bug in the lock mechanism?

BR,

Yordy