From mboxrd@z Thu Jan 1 00:00:00 1970 From: Frank Carreiro Subject: Re: Can't kill a process Date: Tue, 06 Apr 2004 10:47:22 -0600 Message-ID: <4072DF1A.2030608@investools.com> References: <3DF9165145FACB4C96977FF650C1E9040C469D74@its-mail1.its.corp.gwl.com> Mime-Version: 1.0 Content-Type: multipart/alternative; boundary="------------080707070404040509040605" Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com In-Reply-To: <3DF9165145FACB4C96977FF650C1E9040C469D74@its-mail1.its.corp.gwl.com> List-Id: To: "Burnes, James" , reiserfs-list@namesys.com --------------080707070404040509040605 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Yes network connectivity with the box is available. The disk space is slowly dropping. We're running 2.4.9 -e38 on the RedHat Advanced Server. Memory seemed to be fine on the system. Nothing unusually there (1.5 Gigs of ram on the server). I'm recycling the power on the box. Reboot and init 6 seem to run however it's not bringing down the box. Very strange problem :-) Thx Burnes, James wrote: > Do you still have network connectivity with the 10.0.0.5 box? It > sounds like the process is Zombie'ing because the disk subsystem has > lost it's mind or there is some strange network contention race > condition. What version of the kernel are you running? > > > > Have you checked the memory in the destination machine? Reiser is > very demanding of system memory and CPU. It tends to reveal marginal > memory or SDRAM modules that need reseating. > > > > Other than that I think Hans and company may be more help. > > > > Strange. Let me know when you find out what the problem is. > > > > Good luck, > > > > jim burnes > security engineer > great-west, denver > > > -----Original Message----- > From: Frank Carreiro [mailto:frank.carreiro@investools.com] > Sent: Tuesday, April 06, 2004 9:34 AM > To: Burnes, James > Subject: Re: Can't kill a process > > > > I had actually run the following from a foreign system > > tar -czf - .| ssh 10.0.0.5 "cat > /filesystem/files.tgz" > > We were trying to basically maintain the file permissions/ownership > and copy the files to a system with a great deal of diskspace handy. > After a couple of gigs it seemed to stop copying. I had noticed there > is a process 'cat' running and yet I cannot kill it with -15 or -9. > > I did notice that while the files.tgz is not being updated, disk space > is somewhere being eaten up. I'm still poking around to find the > cause. It's a rather strange problem. I've seen this now with two > redhat servers using the reiser file system now. > > Thx > > Burnes, James wrote: > > What is the name of the process? > > > > jim burnes > security engineer > great-west, denver > > > -----Original Message----- > From: Frank Carreiro [mailto:frank.carreiro@investools.com] > Sent: Tuesday, April 06, 2004 9:26 AM > To: reiserfs-list@namesys.com > Subject: Re: Can't kill a process > > > > The system originally was running ext3 until we flattened it and > reinstalled RedHat with the reiser file system. I have other systems > not running reiser which perform just fine. The problem seems to be > pointed at the filesystem > > I had presumed there was something I needed to tweak there. > > > Dave Poirier wrote: > >May I ask how is that related to ReiserFS? Seems like a rather generic Linux problem to me, maybe you could try the Red Hat forums, mailing lists or IRC channels. > > > >Dave Poirier > >Webmestre DSFM / Technicien Informatique > >+1.204.997.6253 > > > >----- Original Message ----- > >From: Frank Carreiro > >Date: Tuesday, April 6, 2004 9:52 am > >Subject: Can't kill a process > > > > > >>I have a rather intense process that has been running since last >> >>night. >> >>I wanted to kill it since it's eating up my cpu however kill and >> >>kill -9 >> >>don't seem to be doing the trick. >> >> >> >>I'm running the reiser 3.6.25 file system on a RedHat Advanced >> >>Server >> >>(2.1) with 2.4.9-3.38smp kernel. >> >> >> >>It's simply counting the time away as it copies files to disk. >> >>I've >> >>reniced it to 15. It's also on a raid 5 array and I've modified >> >>my >> >>min/max speed to 10/50. Performance has improved however the >> >>process is >> >>still running. >> >> >> >>Thoughts? >> >> >> >>thx >> >> >> >> >> >> >> >> >> >> >> > > > > --------------080707070404040509040605 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Yes network connectivity with the box is available.  The disk space is slowly dropping.  We're running 2.4.9 -e38 on the RedHat Advanced Server.  Memory seemed to be fine on the system.  Nothing unusually there (1.5 Gigs of ram on the server).

I'm recycling the power on the box.  Reboot and init 6 seem to run however it's not bringing down the box.  Very strange problem :-)

Thx





Burnes, James wrote:

Do you still have network connectivity with the 10.0.0.5 box?  It sounds like the process is Zombie’ing because the disk subsystem has lost it’s mind or there is some strange network contention race condition.  What version of the kernel are you running?

 

Have you checked the memory in the destination machine?  Reiser is very demanding of system memory and CPU.  It tends to reveal marginal memory or SDRAM modules that need reseating.

 

Other than that I think Hans and company may be more help.

 

Strange.  Let me know when you find out what the problem is.

 

Good luck,

 

jim burnes
security engineer
great-west, denver
 

-----Original Message-----
From: Frank Carreiro [mailto:frank.carreiro@investools.com]
Sent: Tuesday, April 06, 2004 9:34 AM
To: Burnes, James
Subject: Re: Can't kill a process

 

I had actually run the following from a foreign system

tar -czf - .| ssh 10.0.0.5 "cat > /filesystem/files.tgz"

We were trying to basically maintain the file permissions/ownership and copy the files to a system with a great deal of diskspace handy.  After a couple of gigs it seemed to stop copying.  I had noticed there is a process 'cat' running and yet I cannot kill it with -15 or -9. 

I did notice that while the files.tgz is not being updated, disk space is somewhere being eaten up.  I'm still poking around to find the cause.  It's a rather strange problem.  I've seen this now with two redhat servers using the reiser file system now.

Thx

Burnes, James wrote:

What is the name of the process? 

 

jim burnes
security engineer
great-west, denver
 

-----Original Message-----
From: Frank Carreiro [mailto:frank.carreiro@investools.com]
Sent: Tuesday, April 06, 2004 9:26 AM
To: reiserfs-list@namesys.com
Subject: Re: Can't kill a process

 

The system originally was running ext3 until we flattened it and reinstalled RedHat with the reiser file system.  I have other systems not running reiser which perform just fine.  The problem seems to be pointed at the filesystem

I had presumed there was something I needed to tweak there.


Dave Poirier wrote:

May I ask how is that related to ReiserFS?  Seems like a rather generic Linux problem to me, maybe you could try the Red Hat forums, mailing lists or IRC channels.
 
Dave Poirier
Webmestre DSFM / Technicien Informatique
+1.204.997.6253
 
----- Original Message -----
From: Frank Carreiro <frank.carreiro@investools.com>
Date: Tuesday, April 6, 2004 9:52 am
Subject: Can't kill a process
 
  
I have a rather intense process that has been running since last 
night.  
I wanted to kill it since it's eating up my cpu however kill and 
kill -9 
don't seem to be doing the trick.
 
I'm running the reiser 3.6.25 file system on a RedHat Advanced 
Server 
(2.1) with 2.4.9-3.38smp kernel.
 
It's simply counting the time away as it copies files to disk.  
I've 
reniced it to 15.  It's also on a raid 5 array and I've modified 
my 
min/max speed to 10/50.  Performance has improved however the 
process is 
still running. 
 
Thoughts?
 
thx
 
 
 
 
    
 
 
--------------080707070404040509040605--