Sam West | 22 Aug 23:44 2013
Picon

fifo scheduler broken in 4.2.3.1 and above?

Greetings,

I found the following unresolved question on the board:

   http://comments.gmane.org/gmane.comp.clustering.torque.user/13058

We are new users of torque, and just installed version 4.2.4.  Everything appears to
be correctly configured, but when we submit a job, we have the same difficulty as is
documented above, i.e.:

   pbs_statserver failed: 15033
   Problem with creating server data strucutre 

strace output from the pbs_sched process, following a qsub, is:

------
.
.
.
accept(4, {sa_family=AF_INET, sin_port=htons(925), sin_addr=inet_addr("131.10.76.128")}, [16]) = 9
fcntl(8, F_GETFL)                           = 0x2 (flags O_RDWR)
read(8, "\0\0\0\1", 4)                      = 4
rt_sigprocmask(SIG_BLOCK, [HUP INT TERM], [], 8) = 0
alarm(180)
gettimeofday({1377207029, 447005}, NULL) = 0
stat("/etc/localtime", {st_mode=S_IFREG|06454, st_size=327, ...}) = 0
gettimeofday({1377207029, 447640}, NULL) = 0
write(8, "+2+22+21+9Scheduler+0+0+0", 25) = 25
iotl(8, FIONREAD, [0]                    = 0
poll([{fd=8, events=POLLIN|POLLHUP}], 1, 300000) = 1 ([{fd=8, revents=POLLIN|POLLERR|POLLHUP}])
rcvfrom(8, "", 7, MSG_PEEK|MSG_DONTWAIT, NULL, NULL) = 0
write(2, "pbs_statserver failed: 15033\n", 29) = 29
write(2, "Problem with creating server data strucutre\n", 44) = 44
alarm(0)                                       = 180
close(8)                                       = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
.
.
.
--------------                                

Is this a known problem?  Is there a fix?

Thanks.

- s.west
_______________________________________________
torqueusers mailing list
torqueusers <at> supercluster.org
http://www.supercluster.org/mailman/listinfo/torqueusers
Sam West | 23 Aug 17:52 2013
Picon

Re: fifo scheduler broken in 4.2.3.1 and above?

resending, since the original seems to have disappeared into the ether...

From: Sam West
Sent: Thursday, August 22, 2013 2:44 PM
To: Torque Users Mailing List
Subject: fifo scheduler broken in 4.2.3.1 and above?

Greetings,

I found the following unresolved question on the board:

   http://comments.gmane.org/gmane.comp.clustering.torque.user/13058

We are new users of torque, and just installed version 4.2.4.  Everything appears to
be correctly configured, but when we submit a job, we have the same difficulty as is
documented above, i.e.:

   pbs_statserver failed: 15033
   Problem with creating server data strucutre 

strace output from the pbs_sched process, following a qsub, is:

------
.
.
.
accept(4, {sa_family=AF_INET, sin_port=htons(925), sin_addr=inet_addr("131.10.76.128")}, [16]) = 9
fcntl(8, F_GETFL)                           = 0x2 (flags O_RDWR)
read(8, "\0\0\0\1", 4)                      = 4
rt_sigprocmask(SIG_BLOCK, [HUP INT TERM], [], 8) = 0
alarm(180)
gettimeofday({1377207029, 447005}, NULL) = 0
stat("/etc/localtime", {st_mode=S_IFREG|06454, st_size=327, ...}) = 0
gettimeofday({1377207029, 447640}, NULL) = 0
write(8, "+2+22+21+9Scheduler+0+0+0", 25) = 25
iotl(8, FIONREAD, [0]                    = 0
poll([{fd=8, events=POLLIN|POLLHUP}], 1, 300000) = 1 ([{fd=8, revents=POLLIN|POLLERR|POLLHUP}])
rcvfrom(8, "", 7, MSG_PEEK|MSG_DONTWAIT, NULL, NULL) = 0
write(2, "pbs_statserver failed: 15033\n", 29) = 29
write(2, "Problem with creating server data strucutre\n", 44) = 44
alarm(0)                                       = 180
close(8)                                       = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
.
.
.
--------------                                

Is this a known problem?  Is there a fix?

Thanks.

- s.west
_______________________________________________
torqueusers mailing list
torqueusers <at> supercluster.org
http://www.supercluster.org/mailman/listinfo/torqueusers

Gmane