problem with the site_black_list and regular expression in crab jdl

[ericvaandering]

Original Savannah ticket 96325 reported by belforte on Wed Jul 25 10:32:15 2012.

There is a problem with the black list used as default by crab. In the case of dataset stored in the T2_TW_Taiwan , the remote site (the name of ce(s) associated with the t2_tw_taiwan se f-dp001.sinica.edu.tw are f-cream01 and f-cream02) can not be matched because in the jdl the ce associated with the T1_TW are called cream01 and cream02 and the names are contained in the regular expression in order to skip them. But in this way also the ces of T2_TW_Taiwan are skipped.

So a temporary solution could be to remove the T1_W from the black list (file /afs/cern.ch/cms/LCG/crab/config/site_black_list.conf)

Then it is necessary to change the regular expression in the jdl created by crab, adding the /^WORD_TO_MATCH$/ around the name of CE. So the match is done using the correct name and not only a part of it.

The feedback ticket is https://hypernews.cern.ch/HyperNews/CMS/get/crabFeedback/5899/1.html

[ericvaandering]

Comment by belforte on Sat Aug 18 09:16:24 2012

will be fixed for good by using SE names (not regexp) in default black list. Or even better by using only SE b/w list in all places.

[ericvaandering]

Closed by belforte on Thu Sep 12 06:13:03 2013

[ericvaandering]

Comment by belforte on Thu Sep 12 06:13:03 2013

we now only use SE names in default black list. So this is not relevant any more