TM1 Forum

Hi,

Got this issue reported by a customer and it seems that the service got corrupted.
Last saturday without any manual action (at least that's how it's being reported), the server service got disconnected and didn't recover (unfortunately I don't have much more details).

When the customer team tried to start the service (both with cognos configuration or services) they get the error 1067 and it doesn't start.

I've asked them to put the data in a test server and when doing so (or also starting other server services) everything is ok (meaning that for both actions the service started). I'm assuming that the issue is with the service itself.

Based on this and before taking any "strong" action I've decide to write here to check if anything is missing my attention regarding:

- Is there any simple way to solve this apart from killing the server and creating it again.
- I case of doing this, since the admin server has TM1 applications built with performance modeler will I damage anything when killing the service and recreating it again with the same configuration as used in the original tm1s.cfg, server folder and name.

Mainly the second topic is the one I'm more worried due to any link of the TM1 server and the TM1 applications (which are several for this model).

thanks in advance for the help.

Jorge

I presume that a patch had been rolled out to the server and something in that patch has triggered this.

Have you tried creating another tm1 service on the same server and seeing if that starts. If it doesn't then how old a version of TM1 are you on. I had a client on 10.2.2 that had similar problem when it blocked the old ciphersuites. Wouldn't expect this problem to occur if on planning analytics.

If another service does start then 1st thing I'd check is what account is running the tm1 service. Local or named account, could be the named account has expired, credentials changed or permissions changed. If it's set to local you could try assigning a service account to run the service instead. Perhaps the local permissions went awry during a failed Windows update?

I would also recommend trawling through the event viewer to review the events that occurred on Saturday to see if that shows a picture as to what happened.

If you want to regenerate the service for the model you can, just make sure that you know the credentials to any service account set to run it so it can be reapplied later...
To remove a service you can use command prompt with sc delete and the service name. Then open cognos configuration and regenerate it by saving & starting (you shouldn't need to change cognos configuration at all for this test)

Also it would be worth keeping an eye on the resource monitor and double check that the server is running out of memory when starting up (spare ram capacity could differ to the test server). And make sure that the server hasn't ran out of disk space on any drives...

Also if that doesn't help... Could trying removing the tm1s.log, perhap it got corrupted from an unclean termination of the service

Tks.

Log removal was my 1st thinking but it didn't solve it.

Now I'm thinking that something happened to the service itself as the data folder of that server was placed in a test tm1 server/service and it started (also other services start properly).

Based on this I'm thinking on deleting and recreating the service with the same name, tm1s.cfg, ports, location folders, etc.
My only concern is if I damage the tm1 applications created using this tm1 server/service or will it be ok as we will aim to keep everything in the same way.

Kind regards,

Jorge

Issue solved (unfortunately with no clear root cause).

Service must have got corrupted and it had to be rebuilt (delete server in configuration and create a new one with the same name and in the same location - keeping the same ports and config in tm1s.cfg).

Applications were not affected with this action.

Thanks