Huge Archive Log Generation and Slow Data Archive

[Former Member]

Hi Team,

We are Archiving data from our CRM Production Server and during first

few days archiving was smooth and archive log generation at DB level

was also normal, but suddenly from Saturday(08-Feb-2014) process of

slow.

Summary is:

1. Marking file for archival is noraml.

2. Copying data to files is also normal.

3. Deleting Data from DB is taking 6 time more time than normal and

archive log generation during this period for deleting 2500 records is

2.5 GB.

We are looking into this issue using log file and have

found "Checkpoint not complete" which doesnt happen during normal times

but during archival it showing and the solution is to increase log file

size at DB level.

Please let us know, on why system is generating huge archive log and

slow archival during while using TX SARA.

Regards,

RK

[Former Member]

Hi Orkun,

Thanks for your reply ..

We have check memory utilization, substorage performance, VM maybe or database indexes all are normal during process ,

Rableen

[Former Member]

ok. your problem is generating too many redologs correct? Could you check the job at the specific period of time? It may be read tables and create archive files before 08-Feb, not execute the delete job. Because of this, you didn't face with performance problem. After that the delete job has been started on 08-Feb and the system has been started to create more offline redologs during the process.

This is only what I figure out by your responses. To find root cause, it needs to investigate the issue over the system.

Best regards,

Orkun Gedik

[Former Member]

Hi Rableen,

>> We are Archiving data from our CRM Production Server and during first few days archiving was smooth and archive log generation at DB level was also normal, but suddenly from Saturday(08-Feb-2014) process of slow.

There might be many problems cause to the performance issue such as memory utilization, substorage performance, VM maybe or database indexes. Could it be another process interfere and create many offline redologs? You can try to use Oracle log miner to find out what DML statements executed on the specific period of time.

>> We are looking into this issue using log file and have found "Checkpoint not complete" which doesnt happen during normal times but during archival it showing and the solution is to increase log file size at DB level.

This is normal, because delete job drop the records over the table, the system creates many archive logs during the runtime.

Best regards,

Orkun Gedik

[former_member182657]

Hi Rableen,

Please confirm SAP Note   1830577 - Too many logs generated while archiving GTS-specific data

helps you or concerned with your issue or not.

Thanks,

Gaurav

[Former Member]

Hi Nick,

There is no high CPU utilization .it is maximum goes to 40 %.

Rableen

[Former Member]

Hi,

So system is generating huge archive logs and performance (Archiving) is very slow!

If number of archive logs is very high/abnormal, you can observe a poor performance because of multiple log switches.

You can increase the log file size to reduce the number of log switches but it also increases the risk of having  a high data loss (in case of crash).

Regards,

Nick Loy

[Former Member]

Hi Nick,

Thanks for your suggetion , But we are taking only 2500 records and before it was working perfectly .Is there any other suggetion on the same or i have to increse the log size and check??

Rableen

[Former Member]

Hi,

What is the resource utilization, mainly the disk utilization?

Regards,

Nick Loy

[ACE-SAP]

Hello

Nothing special took place ?

Something like an online backup that leave a datafile in backup mode for instance.

That could lead to the same misbehavior.

This could be detected using brconnect -u / -c -f check

Regards