[mdb-discuss] mdb: findleaks: requires UMEM_DEBUG=audit

[xiaolin liu]

Hi Valdemar ,

I catch the memory  thru ::umalog like following ,those was ordered thru time 
stamp. 

Do you know what does mean T-0.000000000 T-0.000006334 ?
Is it the offset from process alive ? what's the unit ? ms ?

or it is the offset from current time ? then the last one (T-11.060110989 ) 
should be the time of process alive?

because I also catch the leak thru TOP command ,which record the system time 
for the leak.

So I want to focus the limited timeframe for  analysis.

Thank you very much .



date time     size
==== ======== =====
12/02 15:31:14 824M
12/03 16:48:08 865M
12/03 16:49:13 872M
12/03 21:05:45 876M
12/04 06:01:18 880M
12/04 10:06:00 881M


::umalog
T-0.000000000  addr=f146000  umem_alloc_8192
         libumem.so.1`umem_cache_free+0x50
         libumem.so.1`process_free+0x78
         AgentSocketPullInput+0xbc
         AgentSocketInputAvail+0xc
         SAPollSubagentEventAux+0x2e4
         __1cWOamCfgInMsgDistributorLmainEvtLoop6F_b_+0x6f0
         __1cLOamCfgMngmtKthreadMain6M_pv_+0x234
         libCpsCmn.so`__1cMCpsCmnThreadHexecute6Fpv_1_+0x68
         libc.so.1`_lwp_start

T-0.000006334  addr=f146000  umem_alloc_8192
         libumem.so.1`umem_cache_alloc+0x210
         libumem.so.1`umem_alloc+0x60
         libumem.so.1`malloc+0x28
         AgentSocketPullInput+0x38
         AgentSocketInputAvail+0xc
         SAPollSubagentEventAux+0x2e4
         __1cWOamCfgInMsgDistributorLmainEvtLoop6F_b_+0x6f0
         __1cLOamCfgMngmtKthreadMain6M_pv_+0x234
         libCpsCmn.so`__1cMCpsCmnThreadHexecute6Fpv_1_+0x68
         libc.so.1`_lwp_start


.....................

T-11.000149406  addr=f146000  umem_alloc_8192
T-11.000155489  addr=f146000  umem_alloc_8192
T-11.000180072  addr=f146000  umem_alloc_8192
T-11.000198906  addr=f146000  umem_alloc_8192
T-11.060057073  addr=f146000  umem_alloc_8192
T-11.060063406  addr=f146000  umem_alloc_8192
T-11.060088239  addr=f146000  umem_alloc_8192
T-11.060110989  addr=f146000  umem_alloc_8192
-- 
This message posted from opensolaris.org