I looked into this and it turns out that condor_router_q is a shell script that does not properly handle JOB_ROUTER_SCHEDD2_POOL when it is a list and not a single item.
if condor_config_val JOB_ROUTER_SCHEDD2_POOL 2>&1 > /dev/null; then
ââââââSECOND_POOL=`condor_config_val JOB_ROUTER_SCHEDD2_POOL 2> /dev/null`
ââââââSECOND_FLAG="$SECOND_FLAG -pool ${SECOND_POOL}"
fi
we would need to change condor_router_q to make this work for you.
I don't think there is any workaround you could make in your config.
-tj
From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> on behalf of Alessandro Italiano <alessandro.italiano@xxxxxxxxxx>
Sent: Wednesday, November 6, 2024 5:06 AM To: HTCondor-Users Mail List <htcondor-users@xxxxxxxxxxx> Subject: [HTCondor-users] condor-ce, JOB_ROUTER_SCHEDD2_POOL list of collectors Hi
I would like to take advantage of the HA feature and use both collector addresses
for the knob
JOB_ROUTER_SCHEDD2_POOL =
htc-ctl-1.recas.ba.infn.it:9618, htc-ctl-2.recas.ba.infn.it:9618
The above conf seems to work, actually jobs are correctly routed to the local pool.
But the same config seems to broke the condor_ce_router_q command which reports the following:
[root@ce-04 ~]# condor_ce_router_q JOBS ST Route GridResource Error: COLLECTOR address or hostname not specified in config file Extra Info: You specified a hostname for a pool (the -pool argument). That should be the Internet host name for the central manager of the pool, but it does not seem to be a valid hostname. (The DNS lookup failed.) 1 [root@ce-04 ~]# condor_ce_q -v $CondorVersion: 23.10.1 2024-10-03 BuildID: 759508 PackageID: 23.10.1-1 GitSHA: acedc362 $ $CondorPlatform: x86_64_AlmaLinux9 $ is it just a matter of sintax ?
Thanks in advance
Ale
|