Scheduling document fails with unknown session (WIS 30567)

building reports webintelligence
1

Hi,

Simple WebI report fails with the following error message…

Could not save the document to the repository for the following reason: [repo_proxy 17] unable to find the info : unknown session (WIS 30567)

Restarted the connection server, still the same error message.

Wireportserver log files has a entry
Could not save the document to the repository for the following reason: [repo_proxy 17] unable to find the info : unknown session (WIS 30567)" PREFIX=“ERR”>
’
Environment : BOE XI 3.1 SP2
CMS : Oracle 10g

Any thoughts…

vijaykollu :us: (BOB member since 2003-05-08)

2

I’m having the same issue…and I’m also on 3.1 SP2. I just applied SP2 a couple days ago…I didn’t have the issue prior, but then it’s inconsistent so there is a slight chance it wasn’t revealed during my scheduling testing, but I don’t think so. Do you know for sure this was introduced with SP2 in your environment?

Skeygo (BOB member since 2006-12-01)

3

I have two environments,

One – 3.1 with fix pack 1.6 (Env1) and other one 3.1 SP2 (Env2)
My scheduling works fine in Env1 but doesn’t work in Env 2…

I raised a support case with BO and they are investigating it…
(they told me that they are not able to replicate it)…

I am tring in different combinations… like moving report from one folder to another / copying reports into anohter folder etc… everywhere same…

vijaykollu :us: (BOB member since 2003-05-08)

4

For me it’s not happening consistently…but about 50% of the time. I’m doing some testing as we speak to narrow it down. Are you connecting to FileStore using UNC path, or is the FileStore on the local server. I have a distributed environment. Two clustered servers, each with full installs. You?

Skeygo (BOB member since 2006-12-01)

5

Yes, mine as well clustered environment and FRS is configured with Fully Qulifier Domain Name (FQDN) path…

Its behaving in-consistent… around 60-40, (3 out of 7 fails)… each time different reports…

vijaykollu :us: (BOB member since 2003-05-08)

6

So I expect you were having the issue before applying any fix-pack/SP2 of the FRS not able to delete objects in the FileStore?

System problem encountered: Failed to remove file \bos01\BOXI31SupportFileStore\FileStore\Output\a_058\150\017\1152570\ from the file system. Please check for permissions or sharing violations

I think it’s possible this issue might get resolved if we can get ahold of a script that will clean up all those files from before that should have been deleted from the file store. Once in a while I get an error saying failed to add an object to the filesstore because it was already there.

Operation failed: Could not save the file \bos01\BOXI31SupportFileStore\FileStore\Output\a_032\163\017\1155872\cdz142c5b25105c4512a.wid to the File Repository Server. File already exists.

Apparently there is a script out there…I’m trying to track it down. I think it’s only a chance the cleanup will fix our problem, but would be good to do regardless.

Skeygo (BOB member since 2006-12-01)

7

Which WAS are you using? I’m using a single webshere on a dedicated server. I don’t think it’s relevant here, but just curious.

Skeygo (BOB member since 2006-12-01)

8

At this client place, I am using Tomcat…

btw, I don’t have any deletion issue before or after applying the SP2…

vijaykollu :us: (BOB member since 2003-05-08)

9

Thanks. I opened a ticket with SAP Friday afternoon with a high priority…hopefully we find a workaround. I’ll let you know.

Skeygo (BOB member since 2006-12-01)

10

Skeygo,

I’m experiencing exactly the same problem as you are with a clustered environment and a filestore connected with UNC path.

I had this kind of message before applying SP2 "System problem encountered: Failed to remove file\ServerName\Input\a_058\150 from the file system. Please check for permissions or sharing violations "

and this one after applying SP2 "Operation failed: Could not save the file \ServerName\Input\a_058\150~cdzkdhdkekdye.wid to the File Repository Server. File already exists. "

So I’m very interested in what the support has to say about it. I’m waiting to see if you post something before opening my own case.

5as :fr: (BOB member since 2009-01-07)

11

Have you checked all your timeouts? The timeouts in tomcat are set to 20 min(?), and the timeouts in BO-XI are less than that. It’s possible that your reportserver session timed out, and your session / the filestore link is broken.

Just a thought, good luck!
B

bdouglas :switzerland: (BOB member since 2002-08-29)

12

That’s a good idea ! Thanks, I’ll try it immediatly !

5as :fr: (BOB member since 2009-01-07)

13

Hi All,

I was getting the same error and what i did was to create a new input and output frs and delete the old ones and it fixed the problem.

Hope this helps everyone else

jte13 :us: (BOB member since 2008-07-09)

14

Thanks…which error did it fix for you…the ‘object already exists’ error, or the failed instance? I deleted and added the SIA on each of my cluster servers, and I was still getting the failed instance issue.

Skeygo (BOB member since 2006-12-01)

15

Delete your AdaptiveJobservers on both nodes and recreate it.

This will fix your scheduling issue.

Thanks
Vijay

vijaykollu :us: (BOB member since 2003-05-08)

16

Thanks Vijay…I’ve tried this already, but it didn’t fix this issue. I would test some more since I think it’s likely it’s not actually fixed in your environment.

Skeygo (BOB member since 2006-12-01)

17

I recreated the FRS and Webi Job servers but it didnt help. :frowning: :frowning:

Sheshachala5 :india: (BOB member since 2004-01-09)

18

As suggested by bdouglas, I increased timeouts for FRS and WebiServer from the initial 20 minutes to 60.

I haven’t seen the issue since. Let’s see if it lasts.

5as :fr: (BOB member since 2009-01-07)

19

I tried the timeouts as well, but the problem is still there. I find the best way to repro it is to schedule the same report 6 or 7 times, for the same time…then ususally at least a couple of them fail.

I’ve noticed something unexpected in the command line of the servers in my cluster. Then -ns switch points to the other server in the cluster.

-fg -restart -name SERVER1.JobServer -ns SERVER2.FQDN:6400 -pidfile “D:\Program Files\Business Objects\BusinessObjects Enterprise 12.0\serverpids\SERVER1_SERVER1.AdaptiveJobServer.pid”

In our R2 clustered environment, the -ns switch would have been the local server (ie. SERVER1 in this example). Not sure if this is related to the problem or not.
Anyone know how to change the default command lines in 3.1?

Skeygo (BOB member since 2006-12-01)

20

We have two servers clustered in this environment. I observed that if the CMS and Webi server doing the scheduling job are on the same server then the schedule is running fine but not if one of it is on the other node.

Sheshachala5 :india: (BOB member since 2004-01-09)