STMS : RFC error: Error when opening an RFC connection

[Former Member]

Hi,

I got a weird issue here, if someone could help:

Our transport (STMS) to other client in domain system, let say client 100 to 200, I getting error * :RFC error: Error when opening an RFC connection*. The same output when I run report RSTPTEST to check the connection.

I tried tp connect from OS level and the return code is OK (0). What may cause this ? Thanks.

Thank a lot

[Former Member]

The tp cannot fork a new process due to saposcol failure.

[Former Member]

Hi Sprout,

dev_col1 is log of saposcol. Its normal size is 50,000 lines.. May be it has exceeded that and is not able to log files , without which it start aborts and therby no forking of process..

try running this commnad:

saposcol u2013ztl

It enables cyclical trace and when dev_col1 is full it writes the log into dev_col2.

When dev_col2 exceeds this size, dev_col1 is deleted and the trace is written to this file again.

Try this more info: http://help.sap.com/saphelp_nw04/helpdata/en/c4/3a6c67505211d189550000e829fbbd/content.htm

Also Try this, it might work ...( if the problem is emerging from shared memory)

as root give ipcs -ma and it will give u many processs running ... check for the one process which ends with dbe , if you see that process give ipcrm -m PID and try to start SAPOSCOL it should work .

Thanks,

Ajith

[Former Member]

Hi,

We had got same error... and we found that tp versions in 2 systems participating were completely different. When we changed it, problem was resolved.

Could you please check the tp versions of the same.

Thanks,

Ajith

[Former Member]

Thanks Ajish,

The tp has the same version. Furthermore the issue was within the domain itself, not with the child system, the issue with tp is done after restarting the system. This no longer be issue.

The only issue left (the cause why tp cannot fork a process) is due to my saposcoll stopped (without me stopping it). SO the issue is now different, when I tried to execute saposcol -l, I got error that dev_coll.tmp cannot be access. In fact, the permission and location is exactly the same with another system whose saposcol running well.

I'll try to issue another message and close this 1 since this is the other issue (not relevant with the subject anymore).

[former_member227283]

Hi,

Manually extent those tablespace and reboot the application once.

Try your activity again.

Thanks,

Anil

[Former Member]

Hi all,

It is not the dbase or filesystem issue. Since the error of tp is cannot fork a process, apparently it is due to memory problem. SAP system cannot allocate OS resource, the saposcol is stopped, the log on ST06 (same in dev_coll) saying

saposcol received stop signalled.

Any ides how this service can be stopped without anyone triggering it ? Thanks.

[former_member227283]

Hi,

which are those 2 tablespace. ???

Thanks

Anil

[former_member227283]

Hi,

ERROR: Could not fork a process: Not enough space

Can you check again , error clearly indicate space problem .

check both the luns is there any space problem :- /usr/sap & /sapmnt

Check any tablespace is full or not

Also do the following acitivty

Goto STMS --> Overview --> system

You will get all your system in present in landscape.

Then select every system one by one and do the below acitivty

Goto Extra menu -- > Generate RFC Destination

Again Goto Extra menu -- > Reset User TMSADM

Do above activity for all system present in STMS

Thanks

anil

[Former Member]

Anil,

I've tried to regenerate and reset TMS user but they are no use.

It is true that there are 2 tablespace are full, by they are autoextend, and sapdata directories are still 80 %....

[former_member227283]

Hi,

Check log , which will describe in which RFC you have problem.

Thanks,

Anil

[Former Member]

Not clear, what log ?

in SMGW the TMSADM status is disconnecting...

The trace file contains error :

• ERROR => GwISearchPartner: timeout, partner LU: >ides<, TP: >/usr/sap/<SID>/DVEBMGS<sys no.>/exe/tp< [gwr3cpic.c 6158]

• ERROR => GwISearchPartner: HOST: ides, conv ID: 66995280 [gwr3cpic.c 6171]

Edited by: sprout009 on Nov 2, 2009 11:03 AM

[Former Member]

Check in SM59 for TMS* RFC Connections. Paste the details.

[Former Member]

Hi Mahesh,

Nothing particular & manual setting in RFC of TMS connection since it is autogenerated, the user ID is TMSADM on client 000 (ABAP connection) both for TMSAMD & TMSUP RFC.

Now the tp connect returns the following error:

ERROR: Could not fork a process: Not enough space

/usr/sap/trans and /tmp is not full

Regards

Edited by: sprout009 on Nov 2, 2009 10:38 AM

[JPReyes]

You must have a full tablespace check DB02

Regards

Juan