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

[HTCondor-users] Explicit specification of execute node locations

Hello Condor Team,Â

We recently tried upgrading fromÂ7.8.8 to 8.2.0, and the following local config file seemingly ceased to work in that no scratch directories were created on any of the D: drive slots under 8.2.0. Â

#Remove Hyperthreading from Blades



#Splits up the cores between the two HDD

SLOT1_EXECUTE =C:\condor\execute

SLOT2_EXECUTE =D:\condor\execute

SLOT3_EXECUTE =C:\condor\execute

SLOT4_EXECUTE =D:\condor\execute

SLOT5_EXECUTE =C:\condor\executeÂÂÂ

SLOT6_EXECUTE =D:\condor\execute

SLOT7_EXECUTE =C:\condor\execute

SLOT8_EXECUTE =D:\condor\execute

SLOT9_EXECUTE =C:\condor\execute

SLOT10_EXECUTE =D:\condor\execute

SLOT11_EXECUTE =C:\condor\executeÂÂÂ

SLOT12_EXECUTE =D:\condor\execute

SLOT13_EXECUTE =C:\condor\execute

SLOT14_EXECUTE =D:\condor\executeÂÂÂ

SLOT15_EXECUTE =C:\condor\execute

SLOT16_EXECUTE =D:\condor\execute

We confirmed that the directories specified above did in fact exist, a necessary condition according to the manual. ÂBecause we have a minimum 2HDs on each machine in our cluster, we were running at about half capacity. ÂRolling back to 7.8.8 did fix the problem, but is somewhat unsatisfactory in that we want to stay up-to-date. Â

Unfortunately my IT help is out this week and we can't install 8.3.0 until he returns next week...so I'm unable to determine if this particular problem persists in 8.3.0. ÂI'm curious to learn if this issue was reported by anyone else when upgrading to 8.2 from 7.x?, and, if not, were there any changes to condor between 7.8 and 8.2 that may have interfered with this functionality?