Re: [Condor-users] can't get Condor-G job to run

["Dr Ian C. Smith" <i.c.smith@xxxxxxxxxxxxxxx>]

--On 06 June 2005 18:03 -0500 Jaime Frey <jfrey@xxxxxxxxxxx> wrote:

> On Jun 6, 2005, at 6:09 AM, Dr I.C. Smith wrote:
>
>
> On Fri, 3 Jun 2005, Jaime Frey wrote:
>
>
> On Jun 3, 2005, at 5:33 AM, Dr Ian C. Smith wrote:
>
>
> --On 02 June 2005 14:50 -0500 Jaime Frey <jfrey@xxxxxxxxxxx> wrote:
>
>
> On Jun 1, 2005, at 8:29 AM, Dr Ian C. Smith wrote:
>
>
> I'm trying to get Condor-G working and I've tried
> submitting an job similar the example in the guide:
>
>
> executable = hello.ksh
> globusscheduler = ulgsmp1.liv.ac.uk/jobmanager-fork
> universe = globus
> output = test.out
> log = test.log
> queue
>
>
>
>
> but it just remains in the idle state. The logfile
> (/tmp/GridmanagerLog.smithic) shows:
>
>
>
>
> 6/1 14:23:43 ******************************************************
> 6/1 14:23:43 ** condor_gridmanager (CONDOR_GRIDMANAGER) STARTING UP
> 6/1 14:23:43 ** /opt1/condor/sbin/condor_gridmanager
> 6/1 14:23:43 ** $CondorVersion: 6.6.7 Oct 11 2004 $
> 6/1 14:23:43 ** $CondorPlatform: SUN4X-SOLARIS29 $
> 6/1 14:23:43 ** PID = 9333
> 6/1 14:23:43 ******************************************************
> 6/1 14:23:43 Using config file: /etc/condor/condor_config
> 6/1 14:23:43 Using local config files:
> /opt1/condor/home/condor_config.local 6/1 14:23:43 DaemonCore:
> Command
> Socket at <138.253.100.177:60984> 6/1 14:23:43 [9333] GAHP server
> pid =
> 9334
> 6/1 14:23:46 [9333] DaemonCore: Command received via UDP from host
> <138.253.100.177:52097> 6/1 14:23:46 [9333] DaemonCore: received
> command
> 60000 (DC_RAISESIGNAL), calling handler (HandleSigCommand()) 6/1
> 14:23:46
> [9333] Found job 142108.0 --- inserting
> 6/1 14:23:46 [9333] Found job 142109.0 --- inserting
> 6/1 14:23:46 [9333] Found job 142110.0 --- inserting
> 6/1 14:23:46 [9333] (142110.0) doEvaluateState called: gmState
> GM_INIT,
> globusState 32 6/1 14:23:46 [9333] (142110.0) proxy not cached yet,
> waiting...
> 6/1 14:23:46 [9333] proxy near expiration or invalid, delaying ping
> 6/1 14:23:46 [9333] (142109.0) doEvaluateState called: gmState
> GM_INIT,
> globusState 32 6/1 14:23:46 [9333] (142109.0) proxy not cached yet,
> waiting...
> 6/1 14:23:46 [9333] (142108.0) doEvaluateState called: gmState
> GM_INIT,
> globusState 32 6/1 14:23:46 [9333] (142108.0) proxy not cached yet,
> waiting...
> 6/1 14:23:46 [9333] GAHP command 'CACHE_PROXY_FROM_FILE' failed:
> Failed
> to import credential maj=851968 min=5 6/1 14:23:46 [9333] ERROR "GAHP
> cache command failed!" at line 357 in file proxymanager.C
>
>
> When I use globus-job-run it's fine so the globus bit seems OK.
>
>
> Any ideas on what it going wrong ?
>
>
> What version of Globus did you create the proxy with, and what
> command
> did you execute (including command-line options)? Globus 4.0
> introduces a
> new proxy format that Condor may have trouble understanding.
>
>
>
>
> I'm still using GT2. I've used this on another host with condor-G and
> it works OK. The command I used was:
>
>
> $ globus-job-run ulgsmp1 -s /home/qcl/smithic/.lfs/condor-g/hello.ksh
>
>
>
>
> I've run a series of globus integration tests and these work OK
> apart from
> gsissh and gsiftp. It look as though condor-g isn't even contacting
> the remote gatekeeper.
>
>
>
>
> Condor-G is failing to acquire its local credentials (i.e. read the
> proxy). Can you turn on D_FULLDEBUG for the gridmanager, try again,
> and post the resulting log?
>
>
>
>
> I've attached the log file. The proxy cert is in /tmp/x509up_u<MY_UID>
> and has read permission just for me - is this OK. AFAIK globus
> complains if you grant read permission for  any other users.
>
>
>
>
> Can you try starting the gahp_server at the command line and typing the
> following line at it:
> CACHE_PROXY_FROM_FILE 1 /tmp/x509up_u41269
>
>
> That is the command that's failing. What does the gahp server reply? If
> the command succeeds, it will print a single 'S'.

No I just get a single 'E' printed.

regards,

-ian.