[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Third-party file staging with Condor-G



On Feb 2, 2008, at 10:41 AM, Jan Ploski wrote:

Jaime Frey wrote:
On Feb 1, 2008, at 2:36 AM, Jan Ploski wrote:

For those who wish to use Condor primarily as a Grid
(meta)scheduler, I see two show-stoppers right now:
1) this third-party transfer feature
2) proxy renewal (and forwarding to the running job) for GT4

Condor does proxy renewal for GT4 GRAM jobs. If you're not seeing
this, please report it as a bug.

I reported problems with that feature a while ago:
https://lists.cs.wisc.edu/archive/condor-users/2007-December/msg00181.shtml

These problems still occur. My test consists of the following steps:
1. myproxy-init -s servername -x -k condor -l root
2. grid-proxy-init -valid 0:20
3. condor_submit myproxy.cmd

myproxy.cmd contains the following lines:

MyProxyHost = servername:7512
MyProxyPassword = ...
MyProxyCredentialName = condor
MyProxyRefreshThreshold = 600

The actual job script sleeps for 30 minutes (during which the proxy used to submit it ceases to be valid) and writes some output to a text file.
The created text file is supposed to be downloaded back to the
submission host after the proxy renewal.

Here is what happens during the test:

1. The GridManager starts running at 16:37:34.

2. The myproxy-get-delegation call to renew the proxy succeeds at 16:47:35.

3. The next interesting events might be these:
2/2 16:54:27 [9274] *** checkDelegation()
2/2 16:54:27 [9274] put_file: going to send from filename
/tmp/condor_g_scratch.0xb38eb0.21197/master_proxy.2
2/2 16:54:27 [9274] put_file: Found file size 5846
2/2 16:54:27 [9274] put_file: sending 5846 bytes

4. The job fails with reason "Globus error: Staging error for RSL
element fileStageOut." In GridmanagerLog, there is also the entry:
http://xml.apache.org/ axis/}stackTrace:org.globus.gsi.proxy.ProxyPathValidatorException:
Certificate C=DE,O=GridGermany,OU=OFFIS e.V.,OU=Grid-RA,CN=Jan
Ploski,CN=201290971 expired.

Can your reproduce these results? Any idea of what is going wrong?


I'd have to see the full gridmanager log to tell whether it's successfully forwarding the renewed proxy to the WS server. Can you give me a copy off-list?

Thanks and regards,
Jaime Frey
UW-Madison Condor Team