On 4/10/02 12:15 PM, "Rob Mayoff" <[EMAIL PROTECTED]> wrote:

>> From Janine's description, it sounds like the "defunct" threads are not
> actually zombies, or else they are not direct children of the nsd
> process that she's killing and their actual parent is not dead.

I was doing what I do with later versions of nsd, which is to kill the
parent process and expect all the children to gracefully die.  I switched to
the "old" version of restart-aolserver, which gets the pids of all the
children and explicitly kills everyone, and that seems to be working better.

Still, I'm restarting the server every 30 minutes and still sometimes find
it effectively hung in between. :(  It doesn't even have to have racked up a
huge number of threads to stop responding.

> Since she's running AOLserver 2.3.3 on Redhat 6.2, this could be rather
> hard to debug.  She said she's using Oracle.  My first action would be
> to tell the Oracle driver to use a TCP or IPC connection to Oracle,
> rather than the "bequeath" adapter, if she hasn't already.  I have seen
> too many problems with the bequeath adapter to trust it for production
> use.

I already do use a TCP connection to Oracle;  we've never used the bequeath
adapter.

This is on 7.2, btw, not 6.2 - on 6.2 it works ok.  On 7.2 it most
definitely does not.  This site ran ok on Solaris, too - it had some
problems there but this was not one of them.

I've got the site programmer working on seeing if we can do a quick 3.3
upgrade, but any suggestions for coping in the meantime will be welcome.
I'm starting to think about doing the restart on a 15 minute interval, which
is getting to be a bit ridiculous!

thanks,

janine

--
Janine Sisk
President/CEO
furfly.net, LLC
Mont Vernon, NH
Phone: 603-672-1122

Reply via email to