Import Monitor hanging - SAPSSEXC

[rainer_fux2]

Hello,

We are doing a migration (based on a complete system export). The system exported was ERP 6.0 (Kernel 720) EHP5 NW 7.2. The database is MaxDB 7.6. The target OS is SuSE Linux 11.

The installation process went very well, but the two last jobs of the import monitor are "hanging". Even after 4 days the jobs did not move forward. And no logs were written. So there cannot be a performance problem.

The jobs are SAPSSEXC (running) and SAPVIEW (waiting). Disk space is okay. No firewall/antivirus is running.

Here is the beginning of the SAPSSEXC.log:

/usr/sap/P13/SYS/exe/run/R3load: START OF LOG: 20111128125609

/usr/sap/P13/SYS/exe/run/R3load: sccsid @(#) $Id: //bas/720_REL/src/R3ld/R3load/R3ldmain.c#13 $ SAP

/usr/sap/P13/SYS/exe/run/R3load: version R7.20/V1.4 [UNICODE]

Compiled Oct 26 2011 21:13:47

/usr/sap/P13/SYS/exe/run/R3load -i SAPSSEXC.cmd -dbcodepage 4103 -l SAPSSEXC.log -nolog -c 50000 -force_repeat -loadprocedure dbsl

(DB) INFO: connected to DB

(DB) INFO: The MaxDB logwriter is switched off#20111128125609

(DB) INFO: Maxdb Kernel version 7.7.06.010

SQLDBC 7.7.6.010

(DB) INFO: AAB_ID_PROP dropped#20111128125609

(SQL) INFO: Searching for SQL file SQLFiles.LST

(SQL) INFO: SQLFiles.LST not found

(SQL) INFO: Searching for SQL file /DVD/MAXDB_4103_LE/ABAP/DB/SQLFiles.LST

(SQL) INFO: found /DVDMAXDB_4103_LE/ABAP/DB/SQLFiles.LST

(SQL) INFO: Trying to open /DVD/MAXDB_4103_LE/ABAP/DB/SQLFiles.LST

(SQL) INFO: /DVD/MAXDB_4103_LE/ABAP/DB/SQLFiles.LST opened

(SQL) INFO: Searching for SQL file SSEXC.SQL

(SQL) INFO: SSEXC.SQL not found

(SQL) INFO: Searching for SQL file /DVD/MAXDB_4103_LE/ABAP/DB/ADA/SSEXC.SQL

(SQL) INFO: /DVDMAXDB_4103_LE/ABAP/DB/ADA/SSEXC.SQL not found

(DB) INFO: AAB_ID_PROP created#20111128125609

(DB) INFO: AAB_ID_PROP~0 dropped#20111128125609

(DB) INFO: AAB_ID_PROP~0 created#20111128125609

(DB) INFO: AAB_ID_PROP deleted/truncated#20111128125609

...

Any ideas?

Best Regards,

[Former Member · Accepted solution]

you can check any longops in database level :

for oracle, adapt it for maxdb:

col SID format B999
col TIME_REMAINING format B999999 heading "Remaining|sec"
col ELAPSED_SECONDS format B999999 heading "Elapsed|sec"
col MESSAGE format A80 WORD_WRAP
set linesize 200
select SID, TIME_REMAINING, ELAPSED_SECONDS, message from v$session_longops
where TIME_REMAINING >0 order by sid;

if you get any output you can see what does current R3load process do.

Regards

Stanislav

Edited by: Stanislav Richnavsky on Nov 28, 2011 3:40 PM

[Former Member]

hi ,

please check the r3 load, database library files which should be latest change it

it may work out

thanks and regards,

santosh

[rainer_fux2]

DC OvFlow: The data cache is operating at capacity and cannot accept 
any new data. Only when pages have been written from the data cache to the
volumes can the data cache accept new pages. If you frequently observe 
DC OvFlow states, check whether your data cache is large enough or the
I/O system too slow.

What can I do to resolve this? As I mentioned I cannot connect to the database (via database manager).

Update:

I had to stop the db because I could not change the parameter (no connection was possible). This way the sapinst stopped as well. I changed the cache size parameter and resumed the sapinst. So far the installation continues. (Two steps of the import monitor are failed. I think because of the already existing TSK file.)

[markus_doehr2]

> (Two steps of the import monitor are failed. I think because of the already existing TSK file.

Execute

R3load -merge_only <TSK-FILE>

in the directory with the TSK files for all those files, who are wrong.

For big migrations I highly suggest

- using migration monitor manually instead of sapinst

- use the newest database version available

- with the newest database version and R3load patches you can use options to speed up the migration:

-merge_bck -para_cnt <number-of-R3load-procesess> -force_repeat -loadprocedure fast

Markus

[rainer_fux2]

Hello,

I have a new problem after I added several more datafiles. The logfiles did not change for hours, and I cannot connect with the database manager to check the database.

The command x_cons P13 sh ac 1 outputs the following:

ID UKT UNIX TASK APPL Current Timeout Region Wait

tid type pid state priority cnt try item

T17 7 9878 Pager Sleeping 0 0 451342(s)

T72 6 9877 GarbCol SVP-End (028) 0 0 272737(s)

T108 4 9875 Savepnt PagerWaitWritr 0 0 324104(s)

T163 8 9879 User 10687* DC OvFlow(046) 0 0 1233892(s)

T187 9 9880 User 14214 DC OvFlow(046) 0 0 3851391(s)

T188 9 9880 User 11436* DC OvFlow(046) 0 0 3851391(s)

T238 11 9882 User 10707 DC OvFlow(046) 0 0 1317(s)

What does that mean? Somehow an overflow?

[rainer_fux2]

Thanks for your help, that should work.

Best Regards,

Rainer

[Former Member]

Hello,

yes your datafiles are full, you can use Database Manager, or Database Studio to add ned datafile.

Regards

Stanislav

[rainer_fux2]

Hello,

We are using MaxDB 7.7.06_10. We use this version because the export was done from this version.

The result of ps -ef|grep R3load was:

p13adm 11525 11491 1 Dec06 ? 00:27:39 /usr/sap/P13/SYS/exe/run/R3load -i SAPSSEXC.cmd -dbcodepage 4103 -l SAPSSEXC.log -nolog -c 50000 -force_repeat -loadprocedure dbsl

p13adm 12354 11491 1 Dec06 ? 00:15:07 /usr/sap/P13/SYS/exe/run/R3load -i SAPAPPL0.cmd -dbcodepage 4103 -l SAPAPPL0.log -nolog -c 50000 -force_repeat -loadprocedure dbsl

root 15968 15715 0 13:18 pts/0 00:00:00 grep R3load

The result of x_cons P13 sh ac 1 was:

ID UKT UNIX TASK APPL Current Timeout Region Wait

tid type pid state priority cnt try item

T82 4 10717 CrIndex DB FULL (197) 0 0 1993835(s)

T108 4 10717 Savepnt DB FULL (198) 0 0 1993835(s)

T163 8 10721 User 11525* DB FULL (197) 0 0 22861091(s)

T188 9 10722 User 12354* DB FULL (197) 0 0 6009469(s

As you can see there are two R3load runs running. I am not shure if I am interpreting the second command right: Does it mean that the database is full?The minimum data size that was needed to beginn the installation was easily excelled. And on the hard disk is enough free space. Any idea whats the problem?

Best Regards,

Rainer

[markus_doehr2]

If the memory (CacheSize) is lower than the biggest index on a table to be created, then database is using temporary pages to create the index. This may lead to a DB FULL as you see it.

Add more datafiles (as already suggested) and the import will go on. You can later then remove the data files again.

Markus

[Former Member]

Hi Rainer Fux,

R3load is looking for SSEXC.SQL file from the following location ((SQL) INFO: /DVDMAXDB_4103_LE/ABAP/DB/ADA/SSEXC.SQL not found). Actually you missed the prerequisite step in the Source system. You have to run SMIGR_CREATE_DDL report in your source system, it will extract the non-stanard data dictonary to a *.SQL files. And then you copy the .SQL files to the following location in your export directory ABAP/DB/ADA/.SQL.

So you have to redo the whole export/import again to fix the issue. And follow the steps in heterogeneous system copy guide.

(SQL) INFO: /DVDMAXDB_4103_LE/ABAP/DB/ADA/SSEXC.SQL not found

Let me know if you still see the issue after redoing the whole export and import.

Regards,

Surendra

[markus_doehr2]

> We are doing a migration (based on a complete system export). The system exported was ERP 6.0 (Kernel 720) EHP5 NW 7.2. The database is MaxDB 7.6. The target OS is SuSE Linux 11.

> The installation process went very well, but the two last jobs of the import monitor are "hanging". Even after 4 days the jobs did not move forward. And no logs were written. So there cannot be a performance problem.

Check using

x_cons <SID> sh ac 1

what the database is doing. What additional options did you use for the import? Recommended is

-merge_bck -loadprocedure fast -para_cnt <number-of-R3load-processes> -force_repeat

Any reason why you use MaxDB 7.6 and not 7.8?

Markus

[former_member709110]

Do you have any R3load process running ?

ps -ef|grep R3load

Regards,

Neel

NB: From the log you have pasted looks like the import was restarted. If that true, why was it done ? and what process was followed?

Edited by: Neelabha Banerjee on Nov 28, 2011 7:15 PM