[GE users] Qmaster hangs frequently (v6.2_u2)

parimi Venkateswara.Rao.Parimi at deshaw.com
Tue Jun 30 19:42:56 BST 2009


If it's in a deadlock do we still expect the threads to report back?

It's not something we could reproduce at will but certainly it's
disturbing for qmaster not recovering from such a deadlock if one exists
even after a restart.

What kind of debugging could be more useful? I've set dl 2 and started
qmaster, which was producing lot of junk as listed below, nothing useful
for a naked eye.

<<>>
4949932  11158 event_master --> sge_event_master_process_send() {
4949933  11158 event_master     Processing event for all clients
4949934  11158 event_master     Preparing event for client 1
4949935  11158 event_master --> eventclient_subscribed() {
4949936  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2571 }
4949937  11158 event_master --> add_list_event_direct() {
4949938  11158 event_master <-- add_list_event_direct()
../libs/evm/sge_event_master.c 2661 }
4949939  11158 event_master     Preparing event for client 11
4949940  11158 event_master --> eventclient_subscribed() {
4949941  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2565 }
4949942  11158 event_master     Preparing event for client 12
4949943  11158 event_master --> eventclient_subscribed() {
4949944  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2565 }
4949945  11158 event_master     Preparing event for client 13
4949946  11158 event_master --> eventclient_subscribed() {
4949947  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2565 }
4949948  11158 event_master     Preparing event for client 14
4949949  11158 event_master --> eventclient_subscribed() {
4949950  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2565 }
4949951  11158 event_master     Preparing event for client 15
4949952  11158 event_master --> eventclient_subscribed() {
4949953  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2565 }
4949954  11158 event_master     Preparing event for client 16
4949955  11158 event_master --> eventclient_subscribed() {
4949956  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2565 }
4949957  11158 event_master     Preparing event for client 17
4949958  11158 event_master --> eventclient_subscribed() {
4949959  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2565 }
4949960  11158 event_master     Preparing event for client 18
4949961  11158 event_master --> eventclient_subscribed() {
4949962  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2565 }
4949963  11158 event_master     Preparing event for client 19
4949964  11158 event_master --> eventclient_subscribed() {
4949965  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2565 }
4949966  11158 event_master     Preparing event for client 20
4949967  11158 event_master --> eventclient_subscribed() {
4949968  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2565 }
4949969  11158 event_master <-- sge_event_master_process_send()
../libs/evm/sge_event_master.c 1663 }
4949970  11158 event_master     processing event master request: 3
4949971  11158 event_master --> sge_event_master_process_send() {
4949972  11158 event_master     Processing event for all clients
4949973  11158 event_master     Preparing event for client 1
4949974  11158 event_master --> eventclient_subscribed() {
4949975  11158 event_master <-- eventclient_subscribed()
../libs/evm/sge_event_master.c 2571 }
4949976  11158 event_master --> add_list_event_direct() {
4949977  11158 event_master <-- add_list_event_direct()
../libs/evm/sge_event_master.c 2661 }
4949978  11158 event_master     Preparing event for client 11
<</>>

------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=38&dsMessageId=204659

To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].



More information about the gridengine-users mailing list