RE: Trace File Question: WAIT_TIME = what? Seconds? Centiseconds?

From: <Christopher.Taylor2_at_parallon.net>
Date: Thu, 11 Apr 2013 11:26:27 -0500
Message-ID: <F05D8DF1FB25F44085DB74CB916678E8879BD8A49B_at_NADCWPMSGCMS10.hca.corpad.net>



Okay, I've had a couple of responses to this and both of them were different :) But it was pointed out that the platform may make a difference and I realized I left that out. Red Hat Linux 5.8 64-bit

I tried cross referencing it with wait times found in 10046 trace events and those are recorded as microseconds I believe but the format for the time portion in a 10046 trace is formatted differently so I wasn't sure if it would be the same here (a DIAG dump / state dump generated by the instance).

Chris

From: Taylor Christopher - Nashville
Sent: Thursday, April 11, 2013 10:40 AM
To: oracle-l_at_freelists.org
Subject: Trace File Question: WAIT_TIME = what? Seconds? Centiseconds?

Just a quick question, in a trace file sample below, what is the wait_time measured in? Clearly it's not seconds unless I have a serious problem...

(I tried finding the answer on Metalink but searching for this specific question is hard to get the results I needed)

    SO: 0x18347e0de8, type: 4, owner: 0x183ede9958, flag: INIT/-/-/0x00     (session) sid: 5321 trans: 0x17f4e15ce0, creator: 0x183ede9958, flag: (40000045) USR/- BSY/-/-/-/-/-

              DID: 0001-008A-000005EE, short-term DID: 0000-0000-00000000
              txn branch: (nil)
              oct: 189, prv: 0, sql: 0x16ee592400, psql: 0x16c4dd9240, user: 92/CONCUITY
    service name: xxxxx
    O/S info: user: xxxxx, term: unknown, ospid: , machine: xxxxxxx

              program: xxxx
    application name: xxxxx, hash value%46894660     waiting for 'enq: TM - contention' blocking sess=0x(nil) seq8 wait_time=0 seconds since wait started16

                name|modeT4d0004, object #;1cd, table/partition=0
    Dumping Session Wait History
     for 'enq: TM - contention' count=1 wait_time24204
                name|modeT4d0004, object #;1cd, table/partition=0
     for 'enq: TM - contention' count=1 wait_timeH9265
               name|modeT4d0004, object #;1cd, table/partition=0
     for 'enq: TM - contention' count=1 wait_timeH9258
                name|modeT4d0004, object #;1cd, table/partition=0
     for 'enq: TM - contention' count=1 wait_timeH9273
                name|modeT4d0004, object #;1cd, table/partition=0
     for 'enq: TM - contention' count=1 wait_timeH9262
                name|modeT4d0004, object #;1cd, table/partition=0
     for 'enq: TM - contention' count=1 wait_timeH9264
                name|modeT4d0004, object #;1cd, table/partition=0
     for 'enq: TM - contention' count=1 wait_timeH9265
                name|modeT4d0004, object #;1cd, table/partition=0
     for 'enq: TM - contention' count=1 wait_timeH9258
                name|modeT4d0004, object #;1cd, table/partition=0
     for 'enq: TM - contention' count=1 wait_timeH9276
                name|modeT4d0004, object #;1cd, table/partition=0
     for 'enq: TM - contention' count=1 wait_timeH9255
                name|modeT4d0004, object #;1cd, table/partition=0
    temporary object counter: 0

Chris Taylor
Oracle DBA
Parallon IT&S

--
http://www.freelists.org/webpage/oracle-l
Received on Thu Apr 11 2013 - 18:26:27 CEST

Original text of this message