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

[Condor-users] Memory leak in condor_procd(.exe)?



Hi,

I'm using Condor 6.9.4 in a testing capacity before I roll it out to all machines and I see that the condor_procd.exe appears to have a memory leak, it is currently using over 52MB of RAM having been active for about 24 hours, and that load is increasing slowly but constantly. Below is a cycle snippet from the StarterLog.slot2 (I use a dual-core machine with 3 slots, one for each CPU and then one for the enitre machine) for when a single CPU is in use.


11/28 13:39:24 In VanillaProc::PublishUpdateAd()
11/28 13:39:24 About to get usage data from ProcD for family with root 1936
11/28 13:39:25 Result of "get_usage" operation from ProcD: SUCCESS
11/28 13:39:25 Inside OsProc::PublishUpdateAd()
11/28 13:39:25 Leaving JICShadow::updateShadow(): success
11/28 13:43:00 Getting monitoring info for pid 2524
11/28 13:47:00 Getting monitoring info for pid 2524
11/28 13:51:00 Getting monitoring info for pid 2524
11/28 13:55:00 Getting monitoring info for pid 2524
11/28 13:55:01 DaemonCore: in SendAliveToParent()
11/28 13:55:01 DaemonCore: Leaving SendAliveToParent() - success
11/28 13:59:00 Getting monitoring info for pid 2524
11/28 13:59:25 Entering JICShadow::updateShadow()
11/28 13:59:25 TokenCache contents:
condor-reuse-slot2@.

Has anyone else seen this behaviour?

Alan