SAP SYBASE IQ Diagnostics and Hang Analysis


 
IQ Diagnostics and Analysis for Hung Server

 
1)      Symptoms

Some of the symptoms, that could lead you (or users) to think that IQ server is hung are :

- No new connections can be made to the server either from the same host machine or different host machine over TCP/IP network.

- No tasks seems to be working and there is no new messages seen in the IQ server log (.iqmsg file), if no existing connections doing work.

- The messages appear to be sparse in .iqmsg file, no task how-ever simple, does appear to get completed in a timely manner.

- Catalog .db and .log time not getting updated on any IQ nodes.

- Hardware bottleneck – IO stuck

 
2)      IQ Diagnostics checklist

Collect and analyzethe following diagnostic information for all IQ problems and/or symptoms

- Know clear problem description and/or exact error message.

- Has anything changed recently on IQ or the OS?

- It is useful to know details like timing of the problem/symptom, duration of the symptom, whether there are any recent changes in the environment.

 
3) DIAGNOSTICS DATA NEEDED TO ANALYZE THE SYMPTOMS

- Analyze the IQ log files - IQMSG.log, IQSRV.log, STDERR.log

- Analyze the Operating System (OS) information: version, patch level, Total memory , and number of CPUs/cores - DSTAT o/p, NMON o/p

- Analyze the OS error logs - /var/log/messages from affected node

- Know if any other tool involved in the problem

- Get Minimum 3 pstack/procstack taken for IQ server PID, 2 to 3 minute apart

- From Coordinator check:

            -- sp_iqmpxinfo

-- sp_iqcontext

            -- sp_iqconnection

            -- sp_iqtransaction

            -- sp_iqstatus

            -- sp_iqversionuse

            -- copy of the IQ cfg file

Comments

Popular posts from this blog