on 05-08-2015 12:02 PM
Hi Experts,
We have upgraded our SAP ECC environment from 6.0 ehp4 to ECC 6.0 ehp7. Before upgrade XXX job was running with 4000 seconds, but post upgrade the same job is running long with 20000 to 25000 seconds.
During the job run, we are getting 900+ exclusive lock entries.
Does this issue related to the SAP Note given below, we are on 740 Kernel release and we are having integrated enqueue WP(not a separate one)
2013043 - Performance Problems with Enqueue Work Process
or should we try some other solution before splitting Enqueue WP from the integrated environment?
Thanks,
Preetha Balan
Hope your kernel version is 740. If it is please upgrade your kernel to latest (if not latest).
There are few performance related issues in the initial level of kernel as we faced during upgrade in our landscape.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi All,
Thanks for your response!!
Our Kernel release is 740. We cannot compare the performance of the job with pervious version as all our systems are upgraded.
Due to these E locks and the long duration taken by this XXX job affects other job's performance and those gets cancelled accessing the same tables as this one.
Is there any other way to increase performance of Enqueue work process to accommodate more/shared locks in order to avoid dead lock situations.
Thanks,
Preetha Balan
Hi Preetha ,
Believe more analysis needs to be done . Having 900 locks at a time is not a very big issue .
Are the other jobs failing because of Enque table lock overflow ? ( Please post the job log of the failed jobs )
If not why do you want to investigate in this manner or what is the reason you believe enqueue is the problem .
Have you taken an ST12 trace for the Job and tried to find out where the issue is actually ?
Thanks ,
Manu
Hello Preetha,
It could be one of many things.
You're on EHP7, what SP level ?
What's your DB ?
Have you analysed where the job is spending it's time ?
If your DB is Oracle, have you scheduled a collect,allsel since the update to EHP7 ?
See what I mean that it could be one of many things ?
Amerjit
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Preeta,
With tx: ST03, check the runtime info of the program. Compare it to the runs before the EHP.
Starting from there, you will see where the difference in runtime is coming from.
It could well be coming from the ENQ as you have considered but ST03 will show you the runtime breakdown which will give you a better idea.
KR,
Amerjit
User | Count |
---|---|
88 | |
23 | |
11 | |
9 | |
8 | |
5 | |
5 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.