topic Re: kernal Parameters in Operating System - HP-UX

kernal Parameters

RKo — Thu, 12 Sep 2002 13:29:55 GMT

I have a two CPC HP9000 Version 11.0.
It has many applications running and these apps are stored different file systems.

At times sytems hangs? Is there any specific kernal params need to be changed?

Anyone has ideas.

Thanks
Reg

Re: kernal Parameters

Sean OB_1 — Thu, 12 Sep 2002 13:46:13 GMT

Reggie,

We'll need more information. What happens when the system hangs? All apps hang? Some apps hang?

What are your current kernel parms?

How much memory, swap, users, etc. use the system?

Anything else you think would be helpful.

Sean

Re: kernal Parameters

Chris Wilshaw — Thu, 12 Sep 2002 13:52:52 GMT

Check memory/CPU/swap use with vmstat, top, swapinfo

Check some number of processes/files with sar -v

Check disk space use with bdf

Check syslog.log, and any application logs for errors

Re: kernal Parameters

Dave Chamberlin — Thu, 12 Sep 2002 13:55:10 GMT

There is not enough info here to answer that question. I think what you need to find out is what resources the system is using and what it spends time waiting for. Without knowing that, changing kernel parameters would be a (risky) stab in the dark. GlancePlus is a excellent tool for seeing what is going on for your system - and you can obtain a trial version. You can also use sar and top to get information. I would check memory resources first. There are many posts in this forum that deal with system resouces and using sar etc to get that information.

Re: kernal Parameters

RKo — Thu, 12 Sep 2002 14:36:20 GMT

Re: kernal Parameters

John Poff — Thu, 12 Sep 2002 14:48:50 GMT

Hi Reg,

It looks like your ttdbserver process is runaway. I just ran into the same problem on a box here this week. The answer was to change an entry in the /etc/inetd.conf file from:

rpc stream tcp swait root /usr/dt/bin/rpc.ttdbserver 100083 1 /usr/dt/bin/rpc.ttdbserver

to:

rpc xti tcp swait root /usr/dt/bin/rpc.ttdbserver 100083 1 /usr/dt/bin/rpc.ttdbserver

Then, do an inetd -c to get inetd to re-read the inetd.conf file, and kill that rpc.ttdbserver process. That should help things short term.

I found that answer by searching in the forum archives. I'd point you to it, but the search engine looks like it is down right now.

JP

Re: kernal Parameters

Anil C. Sedha — Thu, 12 Sep 2002 15:26:23 GMT

Reg,

I agree with John.

You should change your inetd.conf entry for ttdb.server and make it to what he has specified.

Don't forget to assign points. I think John has given you a good answer.

Regards,
Anil

Re: kernal Parameters

RKo — Thu, 12 Sep 2002 16:14:05 GMT

I 'll try short-tem solution. Is there is any longterm solution.

I just checked the syslog.log file and I see FTP was causing the problem??

My swap file size is:

#> vmstat -S
procs memory page faults cpu
r b w avm free si so pi po fr de sr in sy cs us sy id
3 0 0 79948 143400 0 0 0 0 3 0 0 0 -186 540 10 9 81

I am not sure. This is what I see.

Sep 11 15:29:46 h0569n10 telnetd[12696]: getpid: peer died: Error 0
Sep 11 15:30:39 h0569n10 ftpd[12768]: FTP session closed
Sep 11 15:30:46 h0569n10 telnetd[12770]: getpid: peer died: Error 0
Sep 11 15:31:41 h0569n10 ftpd[12971]: FTP session closed
Sep 11 15:31:48 h0569n10 telnetd[12974]: getpid: peer died: Error 0
Sep 11 15:32:42 h0569n10 ftpd[13082]: FTP session closed
Sep 11 15:32:49 h0569n10 telnetd[13084]: getpid: peer died: Error 0
Sep 11 15:33:43 h0569n10 ftpd[13123]: FTP session closed
Sep 11 15:33:50 h0569n10 telnetd[13125]: getpid: peer died: Error 0
Sep 11 15:34:44 h0569n10 ftpd[13181]: FTP session closed
Sep 11 15:34:50 h0569n10 telnetd[13183]: getpid: peer died: Error 0
Sep 11 15:35:45 h0569n10 ftpd[13214]: FTP session closed
Sep 11 15:35:52 h0569n10 telnetd[13216]: getpid: peer died: Error 0
Sep 11 15:36:46 h0569n10 ftpd[13295]: FTP session closed
Sep 11 15:36:53 h0569n10 telnetd[13297]: getpid: peer died: Error 0

Re: kernal Parameters

Sandip Ghosh — Thu, 12 Sep 2002 16:40:20 GMT

Have you done what john and anil had told? What do you mean by long term and short term solution? If you do according to them it will be short term and longterm also. If you want to do it as short term then copy the inetd.conf to someother file and apply the change. If you don't like the change or the problem not resolved then copy back to the original file.

Sandip

Re: kernal Parameters

Tom Maloy — Thu, 12 Sep 2002 16:45:49 GMT

You should set dbc_max_pct and dbc_min_pct. The max defaults to 50%, it should be a % of your RAM, so dbc_max_pct * RAM = 400 MB. And set the min parameter below that. So both might be under 10, depending on how much RAM you have.

Tom

Re: kernal Parameters

RKo — Fri, 13 Sep 2002 11:55:58 GMT

John,

What is the xti stand for? Thats what in your command I see after rpc.

Thanks
Reg

Re: kernal Parameters

John Poff — Fri, 13 Sep 2002 13:40:19 GMT

Reg,

The xti stands for X/Open Transport Interface.

Did you make the change to your inetd.conf yet? If so, did that help things?

JP

Re: kernal Parameters

harry d brown jr — Fri, 13 Sep 2002 13:51:21 GMT

Reg,

Make sure your system patches are up-to-date!

What does

swlist|grep 200[012]

return??

live free or die
harry