nls-archive
[Top] [All Lists]

[nls-archive] Dedicated NLS-restore hardware [was-Re: [cwe-imp] Up2date

To: Jonathan Cheyer <jonathan@cheyer.biz>
Cc: Philip Gust <gust@nouveausystems.com>, "[cwe-imp]" <cwe-imp@cwe.cim3.net>, nls-archive@chm.cim3.net
From: Peter Yim <peter.yim@cim3.com>
Date: Wed, 30 Nov 2005 14:51:22 -0800
Message-id: <af8f58ac0511301451g7f023c1ekee813a26ac148be3@mail.gmail.com>
> [JC] I have not been keeping Ken's KLH10 running on that box in general
> because you had reported that it was running at 99% cpu. In testing on
> my own machine, I have confirmed that it does indeed seem to take up a
> lot of cpu. We may eventually need to dedicate a separate (low-end) box
> for it. That has been the recommendation of both Ken and Mark Crispin.    (01)

[ppy] Yes ... it was a serious processor drain. I don't remember
exactly, I think it was drawing down 50% of the processing power from
each of the two 64 bit 2.4GHz Opteron processors on PASCAL when KLH10
was running.    (02)

Setting up a lower-end dedicated box for it is still on my plate (not
quite near the top, though).    (03)

By the way, do you know if the Museum is still planning to provide the
resources eventually? Not that I mind doing this, but the original
understanding was that we are doing this as a "second source" to what
the Museum was supposed to support. Let's make sure this is still on
Phil Gust's (or the SCC's) to-do list. [attn: Phil]    (04)

Regards.  =ppy
--    (05)


On 11/30/05, Jonathan Cheyer <jonathan@cheyer.biz> wrote:
> Peter,
>
> I have not been keeping Ken's KLH10 running on that box in general
> because you had reported that it was running at 99% cpu. In testing on
> my own machine, I have confirmed that it does indeed seem to take up a
> lot of cpu. We may eventually need to dedicate a separate (low-end) box
> for it. That has been the recommendation of both Ken and Mark Crispin.
>
> For future reference, you can check if KLH10 is running with the ps
> command. It will always be running under the "nls" user. When KLH10 is
> running, you will see processes similar to this listing:
>
> $ ps -ef | grep nls
> nls      14056 14024 99 08:32 pts/5    00:02:03 ./kn10-kl
> ../augment-data/chm.ini
> nls      14057 14056  0 08:32 pts/5    00:00:00 dprpxx -DPM:163841
> nls      14058 14056  0 08:32 pts/5    00:00:00 dprpxx -DPM:196610
> nls      14059 14056  0 08:32 pts/5    00:00:00 dprpxx -DPM:229379
> nls      14060 14056  0 08:32 pts/5    00:00:00 dprpxx -DPM:262148
> nls      14061 14056  0 08:32 pts/5    00:00:00 dprpxx -DPM:294917
> nls      14062 14056  0 08:32 pts/5    00:00:00 dprpxx -DPM:327686
>
>
> Peter, I'll leave it to you to do the reboot and emails.
>
> Thanks!
>
> Jonathan    (06)


> Peter Yim wrote Tue, 29 Nov 2005 23:38:38 -0800:
> > I just ran (remotely) another up2date for both PASCAL (FC4) and MCCARTHY 
>(FC3)
> > again (around 11pm PST Tue 2005.11.29).
> >
> > PASCAL's kernel has been upgraded to  Linux version
> > 2.6.14-1.1644_FC4smp (was: Linux version 2.6.13-1.1532_FC4smp).
> >
> > I did not reboot the box, because I don't know the state of Jonathan's
> > NLS-restore work there. Therefore, Jonathan, please do a reboot when
> > you have a chance. Please confirm when done (with copies to ...[snip]...    (07)
<Prev in Thread] Current Thread [Next in Thread>
  • [nls-archive] Dedicated NLS-restore hardware [was-Re: [cwe-imp] Up2date - PASCAL & MCCARTHY], Peter Yim <=