[HTCondor-users] Groups and SubGroups issue

[Mihai Ciubancan <ciubancan@xxxxxxxx>]

Hello,

I have configured a bunch of groups for my cluster, depending on users and
the number of cores used by their jobs:

RO07AcctGroup = ifThenElse(NordugridQueue =?= "atlas", "group_ATLAS", \
                ifThenElse(NordugridQueue =?= "lhcb",  "group_LHCB", \
                ifThenElse(NordugridQueue =?= "alice", "group_ALICE", \
                ifThenElse(NordugridQueue =?= "ops", "group_OPS" ))))

RO07AcctSubGroup = ifThenElse(regexp("atlas01",Owner) && RequestCpus >1,
"atlas_multicore", \
                   ifThenElse(regexp("atlas01", Owner), "atlas", \
                   ifThenElse(regexp("lhcb01",Owner), "lhcb", \
                   ifThenElse(regexp("pillhcb01",Owner), "pilotlhcb", \
                   ifThenElse(regexp("prdlhcb01",Owner), "prodlhcb", \
                   ifThenElse(regexp("alice01",Owner), "alice", \
                   ifThenElse(regexp("pilalice01",Owner), "pilotalice", \
                   ifThenElse(regexp("ops01",Owner), "ops" ))))))))

AccountingGroup = strcat(RO07AcctGroup, ".", RO07AcctSubGroup, ".", Owner)
ConcurrencyLimits = strcat(RO07AcctGroup, ",", RO07AcctSubGroup, ",", Owner)
SUBMIT_ATTRS = $(SUBMIT_ATTRS), RO07AcctGroup, RO07AcctSubGroup,
AccountingGroup, ConcurrencyLimits

If for Atlas and Alice(as far as I can see) it's working properly, for
LHCb and OPS the mapping is wrong, as you can see from the output of
"condor_status -submitter" command:

Name                         Machine            RunningJobs IdleJobs HeldJobs

alice01@xxxxxxxx             arc6atlas1.nipne.r          36        0        0
atlas01@xxxxxxxx             arc6atlas1.nipne.r           0        0        0
group_ALICE.alice.alice01@ni arc6atlas1.nipne.r         306        1        0
group_ATLAS.atlas.atlas01@ni arc6atlas1.nipne.r          32        7        0
group_ATLAS.atlas_multicore. arc6atlas1.nipne.r         305       79        1
group_LHCB.lhcb.lhcb01@nipne arc6atlas1.nipne.r           0        0        0
group_LHCB.lhcb.pillhcb01@ni arc6atlas1.nipne.r         712      129        0
lhcb01@xxxxxxxx              arc6atlas1.nipne.r           0        0        0
ops01@xxxxxxxx               arc6atlas1.nipne.r           1        0        0
pillhcb01@xxxxxxxx           arc6atlas1.nipne.r           0        0        0
                           RunningJobs           IdleJobs           HeldJobs

    alice01@xxxxxxxx                36                  0                  0
    atlas01@xxxxxxxx                 0                  0                  0
group_ALICE.alice.al               306                  1                  0
group_ATLAS.atlas.at                32                  7                  0
group_ATLAS.atlas_mu               305                 79                  1
group_LHCB.lhcb.lhcb                 0                  0                  0
group_LHCB.lhcb.pill               712                129                  0
     lhcb01@xxxxxxxx                 0                  0                  0
      ops01@xxxxxxxx                 1                  0                  0
  pillhcb01@xxxxxxxx                 0                  0                  0

               Total              1392                216                  1

So the jobs run by user pillhcb01 should be mapped under
group_LHCB_pilotlhcb and not group_LHCB_lhcb, while ops jobs are running
under <none> group instead of group_OPS.

Do you know what I'm doing wrong?

Regards,
Mihai