Workstation hanging

[Mark Stewart]

Hello! I have a j2240 workstation that about once a week seems to hang. Users often leave themselves logged in on workstations on our network and lock it using the screen saver while they are away(it's ok, our network is secluded and secure). This particular workstation seems to hang once a week no matter if many processes or almost none are left running while the user is away. All of the machines use NFS for access to remote drive partitions, but none of these other machines does this. Any suggestions on what might be the cause of this? Thanks!

[Alex Glennie]

When you say hang what exactly is hanging ... for instance the Xserver ie mouse & keyboard or do you mean if you attempt to login via rlogin or telnet no sessions are able to connect or the system does not respond to ping etc ? or combiinations of

I'd toc the system and save the core file for analysis next time this occurs ...what's your O/S and does it have multiple lan cards ?

have you looked around for clues ..syslog.log core file produced ? or run glance etc to gather info on the system performance ?

[Mark Stewart]

By hanging, I mean that the system is not usable. Any processes that are running freeze up and are not able to be killed, mainly because I can't even run top to get the PID of each process. You cannot perform any actions from the prompt, with the exception of "cd", but you can't even get a directory listing. The remote NFSed paritions(on other machines), will not operate as well.

This system is running HP-UX 10.x, and is just a basic workstations setup with a single lan card to communicate with other machines.

As far as I can tell, no core files are produced, but I'll see if toc gives anything, when it happens again, and if I can even get that to work.

Normally, this system runs fine, no problems, no crashes.. but it seems like about once a week, the whole thing just stops. Thanks!

[Uday_S_Ankolekar]

Hi,

Check if any scheduled jobs are running in cron. Is it freezing everyweek during same time..??
-USA..

[Chris Vail]

I'd check to see if the NFS directories were
hard mounted. If they are, then a network
problem, or a reboot of the remote host can
hang the system. If they are, just soft mount them.

[Mark Stewart]

it doesn't crash at the same time each week, it varies, and it doesn't appear that anything is running as a scheduled task..

I'm not sure I follow the difference between hard mounting an NFS directory and soft mounting one. Would you mind elaborating? Thank you!

[John Paine]

Mark,
I had a similar problem on a C class m/c
where it would hang intermittently...turned
out to be the root disk!...made a dd copy
of it whilst m/c was working o.k. and then just swapped the disk out...no more problems since...may not be your problem but worth a shot.
Regards,
John.

[Wodisch]

Hello Mark,

sound like a memory shortage to me: you are running out of usable RAM/swap space...
Check with "swapinfo -t", perhaps even setup a cron-job to execute a couple of times per hour and save the output in a file, so you would be able to check later...

HTH,
Wodisch

[Carl Erhorn]

I've seen this exact set of symptoms when an NFS mount had gone bad. Check each mounted partition, and when you hit the bad one, the system will either hang, or move very slowly.

For instance, an 'ls -la' of a directory was returning a filename about once every 5 seconds. This immediately cleared up once I removed the mount name from /etc/fstab.

However, that is not the way to fix it. As recommended, you will need to edit your NFS mounts so that they fail in soft mode, then reboot the machine so that the mounts are redone. When the bad NFS mount occurs, you will not be able to unmount it, which is why you need to reboot. That gets rid of the bad mount, and will reset the NFS mounts to use the correct mode.

[pap]

Hi,

You should check following things .

1. Check nfs mount points and network connection.If you are hard mounting the nfs file systems and if the network connections gone bad or if you are not able to reach nfs file systems, it may cause to hold your machine. reboot the machine in to single user mod and remove all mounted directories entry from /etc/fstab file.

2. change /etc/rc.config.d/nfsconf and put it NFS_SERVER=0 and NFS_CLIENT=0

Now check the system whether it runs fine.

3. Check the swap memory status. give swapinfo -mt command , if the momory portion is occupied 100% and if it uses some of the device space, then you have to increase swp space/momory.

check whether this machine is under NIS environment. IF it is then may be it hangs if it is unable to reach NIS master server or slave server on its subnet.

Hope it helps.

If at all everything is right, check the machine hardware using stm,xstm etc.

You will get firm decision , what is causing the system hanging problem.

Thanks,
-Piyush.