System very slow

[Swain]

Hi ,

My server is running very slow.
The cause I found out is trans count on devices are more than 100.
could anyone tell what are the other things I should check or how can I track which users are responsible for slowing the system?

running OpenVMS 8.2 on Alpha800 system.1GB RAM

what I found is transaction counts were very high on drives and the que manager was not available. some users logged out and que manager was restarted. Then the system was fine.

The same issue happened the next day. This time trans count was high only.but, queues were available.

let me know what info exactly to be collected in such issues.

Thanks,
Amaresh

[Steven Schweda]

> The cause I found out is trans count on
> devices are more than 100.

Wow. I must be in a world of trouble:

alp $ show dev ALP$Dk

Device Device Error Volume Free Trans Mnt
Name Status Count Label Blocks Count Cnt
ALP$DKA0: Mounted 0 VMS073ALP 12421380 1314 1
[...]

1300 is more than 100, right?

> [...] and the que manager was not
> available. [...]

What, exactly, does this mean?

> let me know what info exactly to be
> collected in such issues.

There's a whole manual on performance
management:

http://h71000.www7.hp.com/doc/os83_index.html
http://h71000.www7.hp.com/doc/73final/6491/6491pro.html

[...]
Preface

This manual presents techniques for evaluating, analyzing, and optimizing performance on a system running OpenVMS.
Discussions address such wide-ranging concerns as:

Understanding the relationship between work load and system capacity
Learning to use performance-analysis tools
Responding to complaints about performance degradation
[...]


As usual, it might help any discussion here
if you provided actual evidence (actual
commands with actual output), rather than
vague descriptions and interpretations.

[Hein van den Heuvel]

It's supposed to run slow... it's a dinky system. :-) :-)

What is the system supposed to do?
What kind of application? Oracle DB? Program development? Email?...
What is it actually doing?
Did it used to run fine, but no more?
What changed, other than the date?

Maybe on other disk (almost) filled up (like in an other topic of yours) and the system is spending excessive time finding disk space?

Start me looking around with the basic MONITOR tool: MONI CLUS (yes, even if it is not a cluster), MONI MODE, MONI PAGE, MONI DISK... MONI PROC/TOPD ... just look around.
$SHOW MEM/FULL
$SHOW MEM/CACH
:

Nothing much going on? Maybe the network slowed to a crawl and is not giving the box enough to do?

If that does not give enough of a clue, maybe try to get T4 or DEecAMDS going, for better insight.

Please report back with an impression of CPU, Memory and disk (IO/sec) usage and perhaps we can guide you onwards.

And uh... Since you failed to provide serious performance characteristics in the initial topic, please consider requesting professional help, as you may not (yet) be ready for the task on hand. There is a lot to learn in this space, but I see no signs that even the basics are understood.

Good luck!
Hein van den Heuvel
HvdH Performance Consulting.

[Steve Reece_3]

You've not really given us enough information to help you properly here.

Is this a change to what you have been experiencing or is this a new installation? If it's a change, what's happened and what's changed? What's different now from when the system was running fine? If it's a new installation, are you expecting the system to do more than it's capable of?

If the transaction counts on drives are higher than you expect (and note the than you expect - every system is different on this) then has an overnight backup failed to shut down one instance of the application yet has started a second one after the backup? Is a disk filling up as other responses to your questions suggest? Is the system disk full (or almost so) and spending time trying to figure out where to put data?
How's the level of fragmentation on disks? Take a large data file on one of your disks and do a DUMP/HEADER/BLOCK=COUNT=0 filename on it. How many headers does the file have? Is this in line with other big files (each header being another file fragment)?

What's "normal" on this system?

The AlphaServer 800 is a nice system but isn't likely to be able to support hundreds of users - it's a small box by present day standards (but then it is around ten or twelve years old).

Steve

[Swain]

Thanks Steven,Hein,Steve!

My apology for not providing quantitative details and disappointing you all with my presence in this forum. The fact is that I am new to OpenVMS. Could anybody guide me getting the best materials for system administration?

Thanks!!

[Hein van den Heuvel]

Your starting point could and should simply be the openvms performance manual.

http://h71000.www7.hp.com/doc/73final/documentation/pdf/ovms_73_perf_mang.pdf

Go play with the SHOW commands and try to figure out what it is trying to tell you. Ditto for monitor. Use HELP.

If I can use only 1 command it would be: MONI MODE. It tell the whole story for those with a keen eye and ear.

Do not focus on the SYSGEN 'tweaks'.
Focus observing and establishing high level issues first:
For example: are you CPU or memory contraint?

Good luck!
Hein.