Troubles with running sapinst (connection reliability)

[symon_braunbaer]

Dear Experts,

this is a very practical question and it's much more OS-related, than SAP-related.

I am performing an action on a 64-bit HP-UX server. I wanted to start sapinst in a

way, that it will be independent on the connection and it won't interrupt its run, if

the connection will be disrupted.

So I have launched:

nohup ../SWPM/sapinst -nogui &

as this is the way, recommended by SAP.

Then, in order to follow the output, I did tail -f nohup.out

The first problem was - when I wanted to do something else in this window, I hit

Ctrl+C and it not only went out of tail, but it also terminated sapinst !!!!!!!!

How come ??? What's wrong with nohup in HP-UX ?

Then I left this window alone and I opened another putty session and started

sapinstgui there.

Now the second problem - I have selected the operation desired on the first screen,

hit next, I could see a couple of lines on the status bar about various .xml files being

generated and so on. Then I saw something like "Determining the SAP systems

installed" and that was it. I never got past that screen. Maybe just one time it went

through, maybe after 25 minutes waiting, but that was all. Once I have hit again

next, the button went grey and it never displayed the next screen.

Did anyone had similar disturbing experience and what could you advise ?

I can only think of requesting VNC next time from the OS guys, as I have been using

it before and it's very reliable.

Many thanks in advance for your suggestions and opinions!

[andreas_herzog]

In  never had problems using SAPinst in the background and following with a tail -f (HP-UX, AIX, SLES...)...why not use at least three sessions/windows? First to start SAPinst, the second for following using tail -f and the third (fourth, fifth whatever) for other actions?

What exactly was shown in the background of the SAPInst process (eg. sapinst_dev.log)?

No action at all? You used ROOT for all installation actions?

GreetZ, AH

[symon_braunbaer]

Hi Andi and many thanks for the reply!

Yes, there is absolutely nothing in the logs. Of course I've used root...

And this weekend it happened again. I was running another DB export, when the connection

got lost, so I decided to give it another try.

I simply run sapinst -nogui (this time I didn't use nohup).

And I have launched sapinstgui in another window.

The export seemed to have resumed. Then I have checked and I was really TERRIFIED with the situation:

- Nothing has happened during the last 2 hours. All the logs were just not being updated for 2 hours !!

- There were 9 R3load processes, COMPELTELY HUNG, I couldn't even get rid of them by using kill -9 !!!!!!

- I had to reboot the server !!!

Please kindly help me to solve this, as it is getting REALLY annoying...

[andreas_herzog]

Very odd...SAPINSTGUI started on Unix or Windows? Did you try both ways?

question: why do you follow SAPINST in a second window with tail -f when you are using a SAPINSTGUI at the same time?

As I already stated...I have not seen such behavior...this problem seems to be a local server problem and not with the SAPINST program itself. We are talking about a "normal" DB export scenario?

I usually start SAPINST with -nogui (starting the GUI on Windows) and leave it for the migration monitor part starting the migmon on UNIX/LINUX command level:

> nohup export_monitor.sh &

and use

> tail -f ExportMonitor.console.log

Any problems with server resources, eg. memory, cpu etc.?

Sorry...no quick solution here

GreetZ, AH

[symon_braunbaer]

Hello Andi,

many thanks for your replies. The advice about running migmon manually on the

command line is a good one. Probably I'm gonna give it a try next time.

For now I'll leave the discussion open, in order to see if someone else won't contribute

a better solution.