swfc@mendelssohn.cs.columbia.edu (Shu-Wie F Chen) (11/09/89)
My NeXT machine has been hanging recently. It seems that my swapfile grows
to my high watermark of 80Mbytes and then thrashes(?) to death. Basically,
I couldn't do anything with my windows (I just get the little disk icon as
the mouse pointer) and I couldn't login from another machine.
Here is what is recorded in /usr/adm/messages...
.
.more of the same
.
Nov 9 11:09:21 mendelssohn mach: vnode_pager: /private/vm/swapfile is full.
Nov 9 11:10:07 mendelssohn mach: vnode_pager: /private/vm/swapfile is available again.
.
.more of the same
.
Nov 9 12:28:31 mendelssohn mach: vnode_pager: /private/vm/swapfile is full.
Nov 9 12:45:37 mendelssohn lookupd[66]: netinfo sleeping: RPC: Timed out
Nov 9 12:45:47 mendelssohn mach: vnode_pager: /private/vm/swapfile is available again.
Nov 9 12:45:47 mendelssohn mach: vnode_pager: /private/vm/swapfile is full.
Nov 9 12:45:47 mendelssohn lookupd[66]: netinfo waking
Nov 9 12:46:10 mendelssohn lookupd[66]: netinfo sleeping: RPC: Timed out
Nov 9 12:48:30 mendelssohn lookupd[66]: netinfo waking
Nov 9 12:49:01 mendelssohn lookupd[66]: netinfo sleeping: RPC: Timed out
.
.more of the same
.
Nov 9 12:58:16 mendelssohn lookupd[66]: netinfo waking
Nov 9 12:58:30 mendelssohn lookupd[66]: netinfo sleeping: RPC: Timed out
I gave up and rebooted here.
What's going on?
Thanks.
Shu-Wie
-------------------------------------------------------------------------------
Shu-Wie F Chen Department of Computer Science
swfc@cs.columbia.edu Columbia Universityavie@wb1.cs.cmu.edu (Avadis Tevanian) (11/11/89)
In article <SWFC.89Nov9104448@mendelssohn.cs.columbia.edu> swfc@mendelssohn.cs.columbia.edu (Shu-Wie F Chen) writes: > >My NeXT machine has been hanging recently. It seems that my swapfile grows >to my high watermark of 80Mbytes and then thrashes(?) to death. >What's going on? Yes, your machine is thrashing to death. You must be running some program with a severe memory leak (unfortunately, virtual memory systems tend to allow these to run for long times). If you notice your swapfile is large, launch a Shell and try something like "ps aux" Look under the VSIZE column (virtual size) for processes with huge virtual sizes. What to do next will depend on what you find. -- Avadis Tevanian, Jr. (Avie) Manager, System Software NeXT, Inc. avie@NeXT.COM
carlson@lance.tis.llnl.gov (John Carlson) (11/14/89)
In article <SWFC.89Nov9104448@mendelssohn.cs.columbia.edu> swfc@mendelssohn.cs.columbia.edu (Shu-Wie F Chen) writes: > >My NeXT machine has been hanging recently. It seems that my swapfile grows >to my high watermark of 80Mbytes and then thrashes(?) to death. Basically, >I couldn't do anything with my windows (I just get the little disk icon as >the mouse pointer) and I couldn't login from another machine. I'm trying to avoid this hanging by (1) setting the hiwater mark (w/ mach-swapon) and (2) NFS mounting another swap file. I'll let you know what happens. John
carlson@lance.tis.llnl.gov (John Carlson) (11/15/89)
In article <613@ncis.tis.llnl.gov> carlson@lance.tis.llnl.gov (John Carlson) writes: >I'm trying to avoid this hanging by (1) setting the hiwater mark >(w/ mach-swapon) and (2) NFS mounting another swap file. I'll >let you know what happens. Here's what I did: (1) Created /etc/swaptab.local (2) added mach_swapon -v -f /etc/swaptab.local -a > /dev/console 2>&1 to /etc/rc.local -- This did not seem to work until I sh'ed /etc/rc.local by hand. Nothing echoed on start up. When I sh'ed by hand, it did echo correctly. (3) Made appropriate changes to /etc/exports file on a Sun. This seemed to work fine for a while (I saw the remote disk light up a few times!), but when I logged in this morning, the system froze when the NeXT dock symbol came up. After trying several mechanisms (Command-Command-`, Command-`, Power Off) I did an Alternate-Left Command *, which rebooted the system. John Carlson carlson@tis.llnl.gov
swfc@mendelssohn.cs.columbia.edu (Shu-Wie F Chen) (11/15/89)
>In article <SWFC.89Nov9104448@mendelssohn.cs.columbia.edu> swfc@mendelssohn.cs.columbia.edu (Shu-Wie F Chen) writes: >> >>My NeXT machine has been hanging recently. It seems that my swapfile grows >>to my high watermark of 80Mbytes and then thrashes(?) to death. >> >>What's going on? > >Yes, your machine is thrashing to death. You must be running some program >with a severe memory leak (unfortunately, virtual memory systems tend to >allow these to run for long times). If you notice your swapfile is large, >launch a Shell and try something like "ps aux" Look under the VSIZE column >(virtual size) for processes with huge virtual sizes. What to do next will >depend on what you find. >-- >Avadis Tevanian, Jr. (Avie) >Manager, System Software >NeXT, Inc. >avie@NeXT.COM Well, we have found the program with the severe memory leak. It seems to be lookupd. Currently, it has a VSIZE of 45.5M, up from 40M of three hours ago. This "phenomena" has also been observed on another NeXT. I had been told by jgreely that this is YP-related and is a known problem. 1. What is the bug/problem 2. When will it be fixed 3. Is there a temporary fix besides rebooting every couple of days or not running YP Thanks. ~swfc ------------------------------------------------------------------------------- Shu-Wie F Chen Department of Computer Science swfc@cs.columbia.edu Columbia University