on 09-23-2014 5:16 AM
Hello Experts,
Today in our Production system, one of the work process has reached 14GB which is surprising.
I couldnt even restart that particulat work process from SM50 as a job is running.
Once the job finished it has moved to ".old" i.e. - dev_w9.old, then i moved to some other directory by which i am able to get some free space
I am still not able to understand how come the file has grown to 14GB though i have the following parameters set:
rdisp/TRACE = 1 (Explanation - 1 Error messages are written to the trace file)
rdisp/TRACE_LOGGING = on, 50 m (Explanation - Trace logging if trace file is larger than 50 MB)
rdisp/autothtime = 60
Could you please let me know of i need to set any more paremeters to restrict the growth of the developer trace file
Thanks,
Subbu
Hi Subbu,
What is this process ? Background ?
Which job was running on this work process? BW or some loading related job ?
On the next run of this, note down the PID on work process and check what is this doing ?
Also, find the ID and then related it to the database using DBACOCKPIT, to find out which query is causing this.
Find out the size of extended + heap for every batch process, that if the limit per process.
May be you can check the logs, what was running on this.
Regards,
Divyanshu
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Thank you very much both.
Yes, that particular process is a background process and it was running a custom report.
When we monitored from Sm50 for this particular process, it was executing some "Z" report and this is not the first time we ran this report.
In the dev_w log of this particular process, it kept writing "ERROR" messages for all work process in the application server:
Actually one background work process, reached the heap limit and all other work process did not consume any heap memory, so it kept writing Error message with the heap memory tlization for all available work process in the application server.:
dev_w log:
A *** ERROR => heap memory WP0: 0 MB [abstor.c 1909]
.
.
.
A *** ERROR => User XXXX reached avail. heapsize = 29344 MB: see ST22 [abstor.c 1904].
.
.
A *** ERROR => heap memory WP<n>: 0 MB [abstor.c 1909]
Is this expected behavior to write continiously about the heap memory if that particular wp is ecpecting heap memory from system?
We are on "Linux 64 Bit operating system, with only SAP application server on it.
It is configured with 63 Gb Memory"
we have below parameters set:
abap/heap_area_dia = 8000683008 - 7.5 GB
abap/heap_area_nondia = 25769803776 - 24 GB
abap/heap_area_total = 30769803776 - 28.5 GB
abap/heaplimit = 40894464 - 40 MB
Please suggest if you recommend any parameter tuning and also please clarify if it is expected behavior to wrire such error information in dev_w log file
Thanks,
Venkat
3 things that you need to do.
Also refer other sections.
Allocating Memory for User Contexts (UNIX) - SAP Memory Management (BC-CST-MM) - SAP Library
2. Refer platform specific memory recommendations. Search for an sapnote and tune it.
3. Improve your zprogram to consume less memory.
Thank you very much Divyanshu.
As part of intial investigation, I had a look at help.sap.com and also referred few SAP notes.
However, I still didnt understand if some other work process used all the heap memory available in the application server, why is this different work process kept writing in the "dev_w" file about the used heap memory information about all the work process?
This basically had caused the file to grow to 14 GB
Coudl you please enlighten me on this topic?
Thanks,
Subbu
Thank you Divyanshu.
We have assigned ABAPers to have a look at Z program.
Trace level - we cant compromise to decrease it as we are logging only errors as of now
one last suggestion needed - I understand below parameter configuration purely depends on the system load and the type of jobs that run on the system.
As mentioned earlier -
We are on "Linux 64 Bit operating system, with only SAP application server on it.
It is configured with 63 Gb Memory"
we have below parameters set:
abap/heap_area_dia = 8000683008 - 7.5 GB
abap/heap_area_nondia = 25769803776 - 24 GB
abap/heap_area_total = 30769803776 - 28.5 GB
abap/heaplimit = 40894464 - 40 MB
By above, please correct me if my understanding is wrong:
1. Each Dialog wp can use up to a max of 7.5 GB and "all DIA wp can use till 28.5 GB" - Am i right?
2. Each BG wp can use up to a max of 24 GB and all BG wp can use a max till 28.5 GB - Am i right?
I am sure i am wrong as one BG used 28.5 GB in this case 😞
So, want to understand the basics of these parameters so that we can start fine tuning based on the jobs that run on the system.
I have gone through the help.sap.com and scn but its still not clear to me
Please clarify
Thanks.
Subbu
Hi Subbu
Could you refer the SAP link
abap/heap_area_total: Total Quota for Heap Memory - SAP Memory Management (BC-CST-MM) - SAP Library
BR
SS
Hi Subbu
Is this long running background job? Could you check the status/any activity going on in transaction SM66?
For Trace parameters you can find the SAP link
Trace Logging - Tools for Monitoring the System - SAP Library
Regards
Sriram
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
85 | |
10 | |
10 | |
9 | |
6 | |
6 | |
6 | |
5 | |
4 | |
3 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.