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

Re: [HTCondor-users] Capital Letters in DNS names



Hi Again,

I should have pointed out that we are still running the emergency patch RC version of HTCondor 9.0.10:

iut2-c300  7:20PM /home/luehring % condor_q -version
$CondorVersion: 9.0.10 Mar 01 2022 BuildID: 576480 PackageID: 9.0.10-0.576480 RC $
$CondorPlatform: x86_64_CentOS7 $


On 4/22/22 7:14 PM, Luehring, Frederick C wrote:
Hi Everyone,

We have a gatekeeper for the US ATLAS site MWT2 that frequently returns its name as iut2-gk.mwt2.ORG rather than the usual iut2-gk.mwt2.org. I find the .ORG form of the address in the startd_history files in all ~250 compute servers at the IU site. If I look at the ClassAds the startd_history file I find both forms:

GlobalJobId = "iut2-gk.mwt2.ORG#717933.0#1649642845"
MyAddress = "<149.165.224.239:9618?addrs=149.165.224.239-9618+[2001-18e8-c02-5-216-3eff-fe72-640e]-9618&alias=iut2-gk.mwt2.ORG&noUDP&sock=shadow_1617_129b_619321>"

and

Environment = "APFCID=13262849.30 APFMON=http://apfmon.lancs.ac.uk/api ATLAS_LOCAL_AREA=/osg/mwt2/app/atlas_app/local OSG_HOSTNAME=iut2-gk.mwt2.org OSG_APP=/osg/mwt2/app HARVE
GridResource = "condor iut2-gk.mwt2.org iut2-condor.mwt2.org:9618"

However it is the .ORG form that may be causing trouble when Panda uses it to form the its own internal batchID using a complicated regex expression. Somehow Panda gets "EPoll: uct2-gk.mwt2.org#92361.0#1650636427" rather than "uct2-gk.mwt2.org#92361.0#1650636427" for itsÂinternal batchID.

Brian Lin (cc'd) helped me in the OSG SW Slack channel and confirmed that:

1) it looks like it's in the schedd name itself
$ podman run --rm -it opensciencegrid/hosted-ce:3.6-release condor_ce_status -pool iut2-gk.mwt2.org:9619 -schedd -af name
iut2-gk.mwt2.ORG

and

2) it doesn't appear that it's coming from config, though

I have grepped extensively in /etc and /var without finding the string ORG in any configuration file though it does of course appear lots in log files and occasionally in files not related to condor.

Thanks for any help!

Fred


--
Fred Luehring Indiana U. HEP mailto:luehring@xxxxxxxxxxx  +1 812 855 1025 IU
http://cern.ch/Fred.Luehring mailto:Fred.Luehring@xxxxxxx +41 22 767 1166 CERN

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature