Hi,

Environment: ASE 15.0.3 ESD#1 on Solaris 10.

In our environments, we have setup a 'admin_sa' Sybase
login ID with all the roles of 'sa' AND a dedicated
temporary database.

During our report runs in the evening, the regular tempdb
is heavily used. At that time, attempts to log into the
ASE server with the 'admin_sa' login id fails with the
following error message:

CT-LIBRARY error:
ct_connect(): user api layer: internal Client
Library error: Read from the server has timed out.

CT-LIBRARY error:
ct_connect(): network packet layer: internal net
library error: Net-lib operation timed out

Observation: The server level optimization goal is set to allrows_dss. At
the time of login failures, the reporting stored procedures
are doing merge_joins and/or hash_joins.

Why does this happen? Are there any additional options
that we can set on the server which will prevent this from
happening?

Thanks in advance.
Anil

1. Nothing in the errorlog?
2. Is dsync off on tempdb devices ?
3. You should setup sysmons during these times, so that you can see where the bottlenecks are.
4. And last but not least upgrade to ESD #4 pls.

Also, note theres a known issue with merge/hash joins in the ESD you are in. I will try to see if I can find the bug #. But know for sure its fixed in ESD #3 and above.

Hi,

Thanks for the response.

No, there is nothing in the error log.

Yes, dsync is off for all tempdbs. Directio is off too for tempdb.

I already ran sysmon during the time. Nothing conclusive.

We are already on the way to upgrading to ESD4#.

If you could dig up the CRs that address this bug in ESD#1, I would really appreciate it. Meanwhile, I will look to see if I can find some reference to it too.

Anil