Hi - hoping someone can please explain to me some behaviour I've noticed within our extreme scale setup. I must however point out that I am not an extreme scale or websphere expert at all. I'm a UNIX sys admin who happens to get involved in all manner of things and network monitoring is one of these things.
We have 4 x Solaris nodes or cache engine members, two in each of our data centres which are located less than 15 miles apart. We have 2 x websphere cells and each cell contains 2 x extreme scale members each. Basically everything is split 50/50 across data centres / cells.
I use ntopng to monitor network traffic by host and protocol. Since extreme scale has been deployed I've noticed that 2 of the extreme scale nodes - 1 in each data centre constantly talk to each other on what seem like random ports and at a rate of around 100 kbytes/sec.
The ports being used are very high, eg. 33364, 33140, 33155, 35606. If I stop / start the extreme scale nodes the ports change but again they are all very high tcp ports as mentioned. So these seem random and are negotiated at startup somehow.
Does anyone know what this might be, our developers have told me that there is nothing configured in either extreme scale or webspehere that uses these ports.
Thanks - Julian.