on 2015 May 14 12:45 PM
Hello all.
We are running a system update on an HA System with two application servers.
The issue we are facing is that when the background jobs scheduled by SUM run on the second Application Server, they fail with an ABAP dump of type RAISE_EXCEPTION with a further explanation of "WRITE_FAILED".
We have identified that the reason this happens is because the background job is trying to write its log under F:\usr\sap\<SID>\SUM\abap\log, and of course that path does not exit on the second App server. When the job executes on the first Application server we have no issues, as this path exists there.
We have gone through all the relevant guides and OSS Notes and we could not find any information on how to handle this issue.
Shouldn't the log be written to a path similar to \\<server_name>\sapmnt\<SID>\SUM... instead of <Drive>:\usr\sap... ?
Is there a workaround to this that we missed in the guides / notes?
Our workaround is to shut down the second App server... which is ridiculous in my humble opinion, but desperate times call for desperate measures, I guess.
Has anyone faced this issue and has a better solution?
Thank you in advance for any help offered.
Request clarification before answering.
Hi Robert,
General recommendation is make DIR_PUT accessible to all the application servers.
Note 1972214 - Error in upgrade :Batch job fails with the error .
Whatever is the DIR_PUT in al11, SUM will write to that and if the mount point is not mounted it will throw error.
DIR_PUT ( Path could be any either /usr/sap/SUM/abap or /sapmnt/SID/SUM any.....) should be mounted on all the application server.
If you dont want to mount /sapmnt/ ** or /usr/sap* check for mount point which is common to all the application server ( could be installation/Media directory) and try to start the SUM tool from there.
Check the space as well in that directory.
Shutting down second app will help but you dont want to do that in Production
Regards
Ranpreet Singh
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
May be /usr/sap/<SID>/SUM dir NFS mounted on PAS and other Dialog instances also.
This would enable SUM background jobs read/write logs on to log dir.
In fact, SUM screen PREP_CONFIGURATION/INITSUBST - suggest this.
Hope this helps.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Robert,
Just a quick turnaround would be shutdown the second application server and let it go or you can try ./STRTUP 'upgdir=,path" as suggested above.
Cheers
Deepanshu Sharma
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Robert,
You can select the instance of your system that shall be used as a background sever but please note that if it is different from your central instance then whatever path you have defined while starting SUM tool should be mounted on that application server.
If you want to choose /sapmnt/SID/SUM as a upgrade directory then you can do that by using option:
./STARTUP "upgdir=<path>"
But make sure that this mount point should be available on application server and enough space should be there as upgrade will create lots of logs.
Another thing is when your upgrade will enter the downtime phase all the background jobs are started on your central instance by default and all the other application servers will be shutdown.
So use of application server can be taken only till downtime phase starts.
Regards
Ranpreet Singh
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.