Recurring jobs overwritten by failed instances

[Naith]

Hi,

I'm running CE85 on a Unix and NT back end (APS on NT, the rest on Solaris. The database is Sybase which isn't fully supported, hence the weird architecture.)

There are some recurring objects which have been set up on the server, which execute once a day perfectly if the report processes successfully.

However, if the instance fails, rather than generate a new failed instance record in the object history, it seems to overwrite the recurring instance - meaning that in 24 hours time, the report will not generate again.

Anybody seen this behaviour before?

Naith

 

[synapsevampire]

ACK!!! Sounds pretty bad, Naith.

Unix isn't the best way to go with CE from everything I've tried and others' horror stories.

Perhaps CD has a hotfix for Unix?

-k

 

[Naith]

No chance, vamp.

Unix isn't as well supported as NT - but as for Unix AND NT? Forget it. It's possible to get around this for v9, as it supports a full Sybase on Unix install.

Thanks anyway,

Naith

 

[rmaich]

My team is experiencing a similar issue on a Windows 2000 installation of CE 8.5. Is there a patch from Crystal that resolves this issue?

Thanks,

Rob

 

[Naith]

It's been a while since this problem came up on that client site, but if memory serves, doing a ps -ef | grep on the crystal services running on the unix box showed some spurious Crystal processes running parallel to the main processes. They looked like duplicates of the original processes, resulting in 20-30 processing jobs instead of the normal 12 or 13.

Taking the Crystal services offline on Unix, bouncing the APS on NT, and then restarting the services on the Solaris platform seemed to rectify the problem.

Hope this helps in some small way,

Naith