remsh

[Jamal Asi]

I'm using remsh on HPUX 11.00 to stop/start SAP
instances on linux application servers. The remsh works sometimes, but some other times it
hangs. The remsh is invoked from a shell script, and I'm using the -n option since I'm using the csh. Any ideas?

[harry d brown jr]

Could your linux servers be running out of available processes/users? Also, is your 11.00 server up to date on patches?

live free or die
harry

[S.K. Chan]

Check this thread ..

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x84d5cf38d6bdd5118ff10090279cd0f9,00.html

[Ron Kinner]

The man says it checks the hostname against either the /etc/hosts or the DNS. Perhaps the DNS is not working/available when your failures occur. I'd check that the /etc/hosts files on the Linux boxes and on your HP are correct and that they are using /etc/hosts before checking DNS.

Ron

[someone_4]

what kind of error do you get when you run the remsh from the command line?

Richard

[Jamal Asi]

I do not get any errors when this is run from the command line. It has worked for me every time I tried it from the command line, and it works from the script but not all of the times. I have three lines in the script that uses remsh to start the SAP instances on the application servers. The first line will start the instance on the first application server. The second line starts the second app server and the third line starts the third app server.What happens when it hangs is that it will start the instance on an application server and then it does not go to the next line to start the other app server or to finish the rest of the script. It could hang on any of the app. servers or it may not hang. The hang occurs after it starts the remote instane. The instance is started fine, but the next line in the script does not get executed and sh is still out there waiting for someting to happen.

[harry d brown jr]

Do you have these patches installed?

PHCO_25707, PHNE_25116, PHNE_25423, AND:
------------

and especially PHNE_23003 where:

SR 8606215328: remshd fails to handle sub shells in some cases.
------------
AND PHSS_21939 where:

The problem did occur because of a change of the remsh behaviour introduced by PHNE_17030. Data can no longer be returned to the calling remsh through stdout or stderr. The SGeSAP is using the remsh to start Application servers. This feature did fail after installing PHNE_17030 because of a syntax change.


live free or die
harry

[Jamal Asi]

I don't have patch PHNE_17030 installed and we are not running SG. HP recommended using 1>&- 2>&- & after the command to cause remsh to return immediately and not wait for the command to complete.

[rick jones]

remsh uses the rather small reserved port number range. this can result in "problems" if a high rate of remsh commands are run (especially if a bunch of rcps or other r-commands are run) it can cause "problems" related to port number reuse and TIME_WAIT and the like.

so, it might be good to see when that second remsh "hangs" what the tcp connection states are on either side with something like netstat -an.

and even better might be to try to get a tcpdump trace of the remsh commands.