problem with TCP/IP please assist

[zoran11]

Hi

I have problem with my scheduled backups, in last few weeks every few days i recive following error in client log on all my clients:


02.09.2005 06:20:21 ANS1017E Session rejected: TCP/IP connection failure
02.09.2005 06:20:21 Will attempt to get schedule from server again in 20 minutes.

I still can ping client and server boxes, but TSM clients can't communicate with TSM server even after i restart them. Only thing that helps is restarting TSM server.
I increased

CommTimeOut 9,600
IdleTimeOut 960

in dsmserv.opt but it did no good.



AIX 5.2
TSM 5.2.2

 

[LED888]

If we are recieving this type of message when we
perform a manual backup or a schedule backup.

ANS1017E Session rejected: TCP/IP connection failure

If we set the following in the dsmserv.opt:
CommTimeOut 9,600
IdleTimeOut 960
Its not going to help since it is not a time out issue.
If it was a time out issue in the actlog there should be a message about 'no communication with node' or 'node have been idle for.'

BTW: If we make changes to the dsmserv.opt file the TSM Server need to be stop/restart to recognize the new update in the dsmserv.opt file. If we are not able to stop/restart the TSM Server check out the 'setopt' command.

Rather using ping, try to 'telnet' or 'ftp' to see if we can establish a connection. If we are not able to establish a telnet or an ftp session , got some kind of network issue.

Also, in the client option file for the tcpserveraddress are we using the fully qualified domain name or the ip address? If we are using the fully qualified domain name try using the ip address of the TSM Server.

Is there a firewall?
If there is are the correct ports open?

Good Luck,
Sias

 

[zoran11]

Thanks Sias

I'll try to be more precise. Telnet,ftp is working just fine everything at AIX or network level is ok, problem is in communication between TSM server and clients (as applications).
Port and IP settings and other settings are also correct (it works for few days and then it simply starts to have that TCP/IP problem and no client can communicate with server).
TSM server can be stoped and started without any problems and then everything is all right for few days...

So if there is any ideas how this could be solved i would appreciate if you could share them with me.

Thanks in advance,
Zoran

 

[LED888]

If the network checks out fine, the TSM Client should connect with the TSM Server with out any problems.

In the dsm.opt (for Windows) or in the dsm.ssy (for Unix), the TCPSERVERADDRESS are we using the ip address or the fully qualified domain name?

If we are using the fully qualifed domain name, use the ip address instead. There may be a problem with the DNS .


Good Luck,
Sias

 

[zoran11]

Thanks again Sias, I appreciate your effort.


I am using IP and there is no network problems, backups are going fine for a period of time(few days), but then TSM Server simply stops communicating with clients (although network between two boxes is OK). When I halt and restart TSM Server then everything is working again.

I know it is difficult to help with this kind of problem (when it works a little bit then stops and then again it works).

Zoran

 

[ProbablyDown]

Is the process still running when failure occurs? what do the logs say?

 

[RMGBELGIUM]

has anything been installed on your system the last few weeks?

rgrds,

R.

 

[zoran11]

The process is still running and server is still listening on port 1500, and nothing was installed in past few weeks.

I still have no idea how to solve this problem and any new input is more than welcome.

Zoran

 

[RMGBELGIUM]

which version of tsm server and tsm client are you using?

rgds,

R.

 

[zoran11]

TSM Server and Clients are 5.2.2

Zoran

 

[ProbablyDown]

Are you behind a router or firewall? Maybe your router is killing sessions after a period of time?