cancel
Showing results for 
Search instead for 
Did you mean: 

Maximum readable joblog on Windows '/usr/sap/<SID>/SYS/global/<clnt>JOBLG'

former_member912992
Participant
0 Kudos

Experts,

There were some error entries in SM21:

.

.

F3V Error 22 for write/read access to a file. File =

F61 TemSe input/output to unopened file

D01 Transaction Canceled TS 004 ( )

F61 TemSe input/output to unopened file

R38 Error at DB commit, return code 001024

.

.

Checking on OS level '/usr/sap/<SID>/SYS/global/<clnt>JOBLG', the contents are not visible on windows explorer. However, listing them in command prompt works.

Earlier we set the job SAP_REORG_JOBS to delete finished and cancelled joblog older than 30 days, but since above case happened, we changed the setting to delete those older than 21 days. Problem solved.

How big is the files limitation for joblogs to be readable by SAP?

Regards,

Agoes

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

Hi,


How big is the files limitation for joblogs to be readable by SAP?

can you explain little bit more what exactly do you want?

regards,

kaushal

former_member912992
Participant
0 Kudos

Hi Kaushal,

When the problem happened, background jobs were failed and there was no log. It seemed failing to write the log. Therefore I suspect there is a limitation of joblogs that can be read / write by SAP.

This is not the first time, we had the issue earlier on another system. The same solution applied.

Regards,

Agoes

Former Member
0 Kudos

Hi,

might be happen only due to Temse inconsistency.

check following sap notes

Note 48400 - Reorganization of TemSe and Spool

Note 12260 - File I/O error when accessing TemSe object

some time it is happens due to faulty kernel try to upgrade your kernel.

if you are facing this problem vary frequently then open OSS message.

regards,

kaushal

former_member912992
Participant
0 Kudos

Hi Kaushal,

According the note 12260, it might be caused by temporary access problem to NFS. So, probably it was caused by intermittent disconnection from app servers to CI. It make sense.

Thanks!

Agoes

Former Member
0 Kudos

We have that same kind of error message but it seems to be caused by online backup of the file system.

Former Member
0 Kudos

please check whether 8.3 short flie names are still created on the volume where the job log files are stored on.

8.3 names are not really needed and the algorithm can run in serious problems if you have several thousand files with similar filenames (that's the case in JOGLG) in one directory.

use command box (cmd.exe)

fsutil -?

fsutil 8dot3name query c:

.....

regards

Peter

Answers (0)