cancel
Showing results for 
Search instead for 
Did you mean: 

CONFLICT_CHECK phase hanging

Former Member
0 Kudos

I installed ERP 6.0 recently and am upgrading to EHP4. My problem is that PREPARE phase CONFLICT_CHECK has been running for over 12 hours already. It looks like it is stuck because none of the log or tmp files are updated. However, when I log into the system I can see a job running in SM50.

It seems that report SAPLOCS_CONFLICTS is doing a sequential read on table OCSCMPLOBJ. This table has 1,565,082 records in it but I don't know what I can do to hurry this along.

I already cancelled the upgrade twice. The first time I just restarted the step. The second time, I stopped and restarted the system and initialized the start of this step.

Can anyone tell me if this is typical for this step? And if it isn't is there something I can do to get this step completed faster?

Thank you for your help.

Peggy

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

Luckily, this is only our Dev system. I expected the entire upgrade to take a few days but didn't expect checking conflicts in SPAM to take this long.

The database is Oracle 10.2.4

Thanks,

Peggy

markus_doehr2
Active Contributor
0 Kudos

That phase checks the new objects vs. the ones being already installed. Depending on the number of packages you have to be imported this may really take a while, however, 12 hours appears to me also very long despite the fact, that all new packages will be checked against all the ones being installed.

You could also check the disk speed and maybe explicitly update statistics on that table again. I'd neverless open an OSS call.

Markus

Answers (4)

Answers (4)

Former Member
0 Kudos

thank you for all the advise. Updating the statistics did the trick. It ran in 2 hours.

Peggy

Former Member
0 Kudos

Thanks. I'll update statistics again and see what happens before opening a message.

Peggy

tomas_black
Employee
Employee
0 Kudos

Hello Peggy,

just a complementary information: this table is initially empty and is filled during the upgrade, so old statistics might result in bad access paths to be calculated from the optimizer, as Markus informed. But what I want to call your attention is that you said you repeated this phase (3 times, right?). So it might be filled with data from previous runs.

My suggestion is that, when you repeat this phase, you choose option (I) from Initial, and not repeat from where it stopped. This will ensure a restart from scratch of this phase, thus likely eliminating unusefull data from this table.

Best regards,

Tomas Black

Former Member
0 Kudos

I ran statistics just prior to starting the EHP4 installation.

I have 8 GB of physical memory with 4GB to the DB. I also have 24GB swap space.

Thanks,

Peggy

markus_doehr2
Active Contributor
0 Kudos

Hi Peggy,

8 GB for an EHP4 installation is VERY few memory, expect a few days of runtime for the installation.

What database are you using?

Markus

markus_doehr2
Active Contributor
0 Kudos

What database are you using and how much memory do you have assigned to it?

Are the database statistics on that table recent?

Markus