Re: [Condor-users] Could not get other processes started except condor_master

[Matthew Farrellee <matt@xxxxxxxxxx>]

Yup, observing that errno 37 is ENOLCK (No locks available) it is 
possible that /home is mounted to not allow locking. A quick search for 
ENOLCK suggests this happens sometimes for NFS mounts.

Best,


matt

On 01/06/2011 12:17 AM, hailong.yang1115 wrote:
> Dear Matthew,
> I have removed all the files under the lock directory
> "/home/863/condor-7.4.2/local.n3/lock/", however the problem remained.
> Any ideas?
> Hailong
> 2011-01-06
> ------------------------------------------------------------------------
> ***********************************************
> * Hailong Yang, PhD. Candidate
> * Sino-German Joint Software Institute,
> * School of Computer Science&Engineering, Beihang University
> * Phone: (86-010)82315908
> * Email: hailong.yang1115@xxxxxxxxx <mailto:hailong.yang1115@xxxxxxxxx>
> * Address: G413, New Main Building in Beihang University,
> * No.37 XueYuan Road,HaiDian District,
> * Beijing,P.R.China,100191
> ***********************************************
> ------------------------------------------------------------------------
> *发件人：* Matthew Farrellee
> *发送时间：* 2011-01-06 06:52:59
> *收件人：* Condor-Users Mail List
> *抄送：* hailong.yang1115
> *主题：* Re: [Condor-users] Could not get other processes started except
> condor_master
> On 12/31/2010 07:22 AM, hailong.yang1115 wrote:
>  > Dear all,
>  > When I started the condor_master process, no other processes were
>  > started and the condor_master exit after a while. I also got the
>  > following error in the MasterLog:
>  > 12/31 19:51:43 ******************************************************
>  > 12/31 19:51:43 ** condor_master (CONDOR_MASTER) STARTING UP
>  > 12/31 19:51:43 ** /home/863/condor-7.4.2/sbin/condor_master
>  > 12/31 19:51:43 ** SubsystemInfo: name=MASTER type=MASTER(2)
> class=DAEMON(1)
>  > 12/31 19:51:43 ** Configuration: subsystem:MASTER local:<NONE>
> class:DAEMON
>  > 12/31 19:51:43 ** $CondorVersion: 7.4.2 Mar 29 2010 BuildID: 227044 $
>  > 12/31 19:51:43 ** $CondorPlatform: I386-LINUX_RHEL5 $
>  > 12/31 19:51:43 ** PID = 20607
>  > 12/31 19:51:43 ** Log last touched time unavailable (No such file or
>  > directory)
>  > 12/31 19:51:43 ******************************************************
>  > 12/31 19:51:43 Using config source:
> /home/863/condor-7.4.2/etc/condor_config
>  > 12/31 19:51:43 Using local config sources:
>  > 12/31 19:51:43 /home/863/condor-7.4.2/local.n3/condor_config.local
>  > 12/31 19:52:13 FileLock::obtain(1) failed - errno 37 (No locks available)
>  > 12/31 19:52:13 ERROR "Can't get lock on
>  > "/home/863/condor-7.4.2/local.n3/lock/InstanceLock"" at line 956 in file
>  > master.cpp
>  > The file InstanceLock indeed existed and had the following permission:
>  > -rw------- 1 863 863 0 May 30 2010 InstanceLock
>  > Any clue?
>  > Hailong
>  > 2010-12-31
> The condor_master maintains an "instance lock" to prevent you from
> accidentally running it multiple times. It may be that the master exited
> and didn't clean up the lock. If you have no master running, try
> removing the file manually.
> Best,
> matt