on 11-13-2012 9:38 PM
Hi Everyone,
I am getting the following error when running the job.
Cannot write onto Disk for the file <D:\Program Files (x86)\SAP BusinessObjects\Data
Services\log\pCache/ptbod1__bodq/924321CacheSplit3_6108_1>. OS Error: <No space left on device>
In the monitor log I see that this particular dataflow is pulling almost 4.8 million records.
DF_CUST_V /Query | PROCEED | 4908000 | 750.639 | 1960.385 |
Also, very interesting thing is we have our production environment clustered (
4 job serveres in separate machines) and we are facing this error only while the job runs on 2 particular jobservers. I cannot really ask the developer to change the design as the job runs perfectly fine on the other 2 servers. As the environment is laod balanced this job can be kicked off from an one of the servers.Few more details about the job:
Thanks,
Pramod
Hi Pramod,
You can concider changing the log directory and pcache of the 2 job servers that are facing disk space problems to a another disk/share with plenty space.
To do this see:
https://service.sap.com/sap/support/notes/1265817
https://service.sap.com/sap/support/notes/1372924
regards
Nawfal
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Nawfal,
Thanks for your help. Creating disk space has worked for us. I see that 40gb free space works for us and I observed the pcache folder size increases when the particular dataflow is in progess.
However, I overlooked this as in the other server the disk space was lesser and the job was able to complete without any issues. I will hopefully find the answer for it.
Thanks for your help, I appreciate it
Pramod
Hi Everyone,
I am getting the following error when running the job.
Cannot write onto Disk for the file <D:\Program Files (x86)\SAP BusinessObjects\Data
Services\log\pCache/bods_repo_admin/2621Sort1528_3604_111>. OS Error: <No space left on device>
Can anyone please help
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Pramod,
I am also getting the same issue which is you got when triggere the BODS validation job after pacting 4.0 to 4.2.
When we place a file in share folder path and triggered the job then the file data moved to BODS staging but the job got error that OS Error:<No Space left on Device>.
Could you please let me how could you resolve your issue.
Please give any suggestions on the error.
Thanks,
Subbu
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
I have tried using the inmemory cache for the dataflow which is causing the issue and it seems like it was using the whole CPU memory ( which is 28 gb ). The server got hanged because of this. I have forced to run the job on the server which it is running without any issues for now.
My next try will be increasing the pageable buffer size to 16 GB and see if that works. As I will be away this sunday for thaksgiving, i will be working on this issue fortnight later.
Do give me the suggestions in between.
Thanks,
Pramod
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Pramod,
What about the files system:
NTFS vs. Fat32 (4GB)
Did you set quotas on NTFS for the user?
Regards
Norbert
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Norbert,
There are no set quotas on the system.
@Everyone:
I will try doing the follwoing:
I will be working on these on every sunday from now ( for each change ) and let you guys know if it worked.
but I still have this question " why is this job running on other jobservers which have less RAM and disk space?"
In the monitor log I see that this dataflow is taking 650s to process 4.8 million rows on the servers it is able to run and on the particular server I am trying to troubleshoot it takes 750-800s. ( I dunno if this would help any one, but this is the observation for now )
Adding one more point,
All the production boxes have a common database. So, the settings and the job are exactly same as far as I checked.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
87 | |
10 | |
10 | |
10 | |
7 | |
6 | |
6 | |
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.