Informix 7.24UC8 performance
Folks,
We have an active application running on Informix 7.24UC8. Our system is
an online system that runs 24 by 7 and we have about 2000 users at any
given time of the day. It has been running without an interuption for 60
days (since Christmas), but yesterday the system just got slower and
slower. Just a query into the sysmaster tables sometime it takes serveral
minutes. After a couple of hours investigation without pinpointing what
may cause the problem, we recycled the database server and everything went
back to normal.
Does anyone have any idea what may cause this degradation problem. We
collected onstat -a (140 MB) prior the shutdown if anyone wants to analyze
it.
Thanks,
ben.truong@amd.com
In article <01bf8489$a3ba1310$18e9b5a3@btruong>, Ben Truong
<Ben.Truong@amd.com> writes
>Folks,
>
>We have an active application running on Informix 7.24UC8. Our system is
>an online system that runs 24 by 7 and we have about 2000 users at any
>given time of the day. It has been running without an interuption for 60
>days (since Christmas), but yesterday the system just got slower and
>slower. Just a query into the sysmaster tables sometime it takes serveral
>minutes. After a couple of hours investigation without pinpointing what
>may cause the problem, we recycled the database server and everything went
>back to normal.
>
>Does anyone have any idea what may cause this degradation problem. We
>collected onstat -a (140 MB) prior the shutdown if anyone wants to analyze
>it.
>
Which platform>
Is NOAGE set to 1 on ONCONFIG?
Check release notes under $INFORMIXDIR/release/en_us/0333/IDS*
>Thanks,
>ben.truong@amd.com
--
David Williams
In article <01bf8489$a3ba1310$18e9b5a3@btruong>,
"Ben Truong" <Ben.Truong@amd.com> wrote:
> Folks,
>
> We have an active application running on Informix 7.24UC8. Our system
is
> an online system that runs 24 by 7 and we have about 2000 users at any
> given time of the day. It has been running without an interuption for
60
> days (since Christmas), but yesterday the system just got slower and
> slower. Just a query into the sysmaster tables sometime it takes
serveral
> minutes. After a couple of hours investigation without pinpointing
what
> may cause the problem, we recycled the database server and everything
went
> back to normal.
>
> Does anyone have any idea what may cause this degradation problem. We
> collected onstat -a (140 MB) prior the shutdown if anyone wants to
analyze
> it.
>
> Thanks,
> ben.truong@amd.com
>
There are several possibilities. You may have started allocating more
shared memory and gotten it fragmented. Check ipcs -a and look for
informix or onstat -g mem, I believe.
You may also have run into a problem where the operating system reduces
the priority of the oninit process because it had been running for so
long. Check your release notes and see if NOAGE in your config file is
supported.
These are a couple of ideas. I'm sure others will have more ideas.
--
# unrm /
ksh: unrm: not found
# man cpio
Sent via Deja.com http://www.deja.com/
Before you buy.
Check your onstat -R, sometimes BUFFERS will get loaded with index pages such
that users then compete for placing data pages. There is an environment
variable that can be set to prevent index pages from having a higher priority
over data pages - call tech support for the name of the variable.
Nona