[OpenAFS-devel] Re: [OpenAFS] OopenAFS 1.2.13: ever increasing number of fileserver connections - h_Hold leak

Renata Maria Dart Renata Maria Dart <renata@slac.stanford.edu>
Thu, 17 Feb 2005 08:51:26 -0800 (PST)


Hi, I applied Rainer's fix to a lightly loaded solaris 9 fileserver and I am
still seeing entries in the FileLog that look like:

Thu Feb 17 08:35:02 2005 FindClient: client ea7e58(7982e684) already had conn 
eaa790 (host 808a850c), stolen by client ea7e58(7982e684)
Thu Feb 17 08:40:02 2005 FindClient: client e7a550(7982e6b8) already had conn 
eaa930 (host 808a850c), stolen by client e7a550(7982e6b8)
Thu Feb 17 08:45:02 2005 FindClient: client ea69c8(7982e6ec) already had conn 
ea81a0 (host 808a850c), stolen by client ea69c8(7982e6ec)

Is this an indication that the problem still exists and if I load
the fileserver up I will again see the ever increasing host/client connections
problem?

Thanks,

Renata


>X-Filter: CERNMX03 SMTPGW CERN Spam Sink v1.0
>From: Rainer Toebbicke <rtb@pclella.cern.ch>
>To: openafs-info@openafs.org
>Cc: openafs-devel@openafs.org
>Subject: Re: [OpenAFS] OopenAFS 1.2.13: ever increasing number of fileserver 
connections - h_Hold leak
>X-OriginalArrivalTime: 07 Jan 2005 11:17:16.0546 (UTC) 
FILETIME=[71E1D620:01C4F4AA]
>X-Virus-Status: Clean
>X-Greylist: Delayed for 00:12:36 by milter-greylist-1.6rc1 
(meredith.dementia.org [128.2.120.216]); Fri, 07 Jan 2005 06:30:13 -0500 (EST)
>List-Post: <mailto:openafs-info@openafs.org>
>List-Subscribe: <https://lists.openafs.org/mailman/listinfo/openafs-info>, 
<mailto:openafs-info-request@openafs.org?subject=subscribe>
>List-Unsubscribe: <https://lists.openafs.org/mailman/listinfo/openafs-info>, 
<mailto:openafs-info-request@openafs.org?subject=unsubscribe>
>List-Archive: <https://lists.openafs.org/pipermail/openafs-info/>
>Date: Fri, 07 Jan 2005 12:17:15 +0100
>
>Rainer Toebbicke wrote:
>> Rainer Toebbicke wrote:
>> 
>>> Something's wrong with OpenAFS 1.2.13 as on several busy servers we 
>>> see an ever-increasing number of host/client connections.
>>>
>
>The attached patch (against 1.2.13, but modulo line numbers also for 
>1.3.73 and later) fixes a h_Hold table leak in GetSomeSpace_r:
>
>lih_r now leaves the current lih_host held, and h_Releases those for 
>which it changed mind during h_Enumerate. It also closes the window that 
>ClearHostCallbacks_r opens.
>
>The problem was: due to the leak on the h_Hold table, rx connections and 
>host and client structures stopped being garbage collected once the 
>fileserver went through GetSomeSpace_r. Only relevant for "busy" 
>servers, many would never even invoke this routine. The problem did not 
>appear before 1.2.11 as the delta that caused this was not yet in.
>
>Tested on two servers up to now.
>
>-- 
>=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
>Rainer Toebbicke
>European Laboratory for Particle Physics(CERN) - Geneva, Switzerland
>Phone: +41 22 767 8985       Fax: +41 22 767 7155

 Renata Dart                         | renata@SLAC.Stanford.edu  
 Stanford Linear Accelerator Center  |    
 2575 Sand Hill Road, MS 97          | (650) 926-2848 (office)
 Stanford, California   94025        | (650) 926-3329 (fax)