tracking a pid

[mike]

hi

i got an error message that say

genunix: [ID 470503 kern.warning] WARNING: Sorry, no swap space to
grow stack for pid 19651 (Cgistub)
Any ways to track down this pid 19651 as it happens? Enabling BSM
would require a restart of my server which is not feasible . can
enabling accounting do the job? what other ways can i do it?
thanks

[Barton Fisk]

On Sat, 26 Jul 2003 08:44:59 -0700, <"Michael Vilain <vilain@spamcop.net>">
wrote:

> In article <dfd17ef4.0307250051.56027bf8@posting.google.com >,
> [email]s99999999s2003@yahoo.com[/email] (mike) wrote:
>

>>
>> genunix: [ID 470503 kern.warning] WARNING: Sorry, no swap space to
>> grow stack for pid 19651 (Cgistub)
>> Any ways to track down this pid 19651 as it happens? Enabling BSM
>> would require a restart of my server which is not feasible . can
>> enabling accounting do the job? what other ways can i do it?
>> thanks

>
> Sorry. You'll have to schedule a reboot. If you're out of swap space,
> you'll have to install an additional swap file and add it to vfstab.
> Reboot to verify the vfstab entry is OK and the system is in a stable,
> known state.

hmmmm...

swap -a works for me. no reboot needed.

--
Barton Fisk

Using M2, Opera's revolutionary e-mail client: [url]http://www.opera.com/m2/[/url]

[Barton Fisk]

On Sat, 26 Jul 2003 20:31:43 -0700, <"Michael Vilain <vilain@spamcop.net>">
wrote:

> In article <oprsx58rfr45l1c8@newsgroups.bellsouth.net>,
> Barton Fisk <bfisk7@bellsouth.net> wrote:
>

>> On Sat, 26 Jul 2003 08:44:59 -0700, <"Michael Vilain
>> <vilain@spamcop.net>"> wrote:
>>

>> > In article <dfd17ef4.0307250051.56027bf8@posting.google.com >,
>> > [email]s99999999s2003@yahoo.com[/email] (mike) wrote:
>> >
>> >>
>> >> genunix: [ID 470503 kern.warning] WARNING: Sorry, no swap space to
>> >> grow stack for pid 19651 (Cgistub)
>> >> Any ways to track down this pid 19651 as it happens? Enabling BSM
>> >> would require a restart of my server which is not feasible . can
>> >> enabling accounting do the job? what other ways can i do it?
>> >> thanks
>> >
>> > Sorry. You'll have to schedule a reboot. If you're out of swap

>> space, > you'll have to install an additional swap file and add it to
>> vfstab. > Reboot to verify the vfstab entry is OK and the system is in
>> a stable, > known state.
>>
>> hmmmm...
>>
>> swap -a works for me. no reboot needed.

>
> Yes, but suppose you mistyped the entry in /etc/vfstab. It's happened.
> Then you won't come up cleanly. At the next power fail or downtime at
> 2am, I'd rather not get a phone call from Operations.
>
> It sounded like the OP was in a production environment where downtime is
> frowned upon. Just doing a "swap -a" constitutes a system change that
> must be verified, in my book, to ensure the system can boot after the
> change.
>
> This may sound foreign and overly constraining, but that's what real
> production environments (i.e. ones that measure downtime using 3-9's 5 or
> better [about 21.6 minutes/month]).
>

I'm aware of varying production requirements and change control policies.
My point was that no reboot is required to add swap space. It can be done
on the fly (swap -a) and verified at a scheduled outage window.

Also, you are assuming that the swap space increase will be permanent. This
may not be the case after the application issues are identified.

Never assume in a production environment.

--
Barton Fisk

Using M2, Opera's revolutionary e-mail client: [url]http://www.opera.com/m2/[/url]