Is the opening complaint regular or does it happen at particular times of the day ?
CONSTANT - Gather a couple of statspack reports for tiny durations of time when the complaint occurs CERTAIN TIMES - Gather a statspack report for a time of time when the complaint exists.
ADDITIONALLY accumulate a statspack report for a identical time of time when the complaint does not happen for comparison.
NOTE:- As sufficient as probable statspacks reports should be minimum 10 minutes, limit 30 minutes. Longer durations can crush the infomation and reports should be re-gathered using a shorter time period.
Does the complaint start a session, a few sessions or all sessions ?
ONE SESSION - Gather 10046 keep track is to session. SEVERAL SESSIONS - Gather 10046 keep track for a or two of the complaint sessions ALL SESSIONS - Gather statspack reports Does the database ''actually'' cling to or only ''appear'' to hang?
(ie do sessions never total their tasks (HANG or SPIN?) or do they it finally finish (SLOW) ) HANG - Take some systemstates together with a statspack report SPIN? - See:Â Document 68738.1 No Response from the Server, Does it Hang or Spin? SLOW - Gather 10046 for a preference of slow sessions.
Is the CPU use high for a or more sessions when things run bit by bit ? YES - Take some errorstacks from the think CPU process. (* If not able to to accumulate errorstacks then accumulate pstack reports)
Note: If uploading reports to support, greatfully make sure that they are in Text format For 9i and 8i, statspack, rsther than than AWR, reports should be gathered. To accumulate a statspack report, greatfully impute to Document 94224.1 FAQ- Statspack Complete Reference.
To appreciate statspack outlay impute to: />
Login to sqlplus as the inner user: sqlplus "/ as sysdba"
login to SQL*Plus:
% book pstacks.txt % /usr/proc/bin/pstack pid % exit
The PID is the o/s routine id of the routine to be traced. Repeat the pstack order about 10 times to takeover probable smoke-stack changes. Further sum of pstack are in Document 70609.1 How To Display Information About Processes on SUN Solaris PLSQL Profiler.
The PL/SQL profiler provides data abour PL/SQL ethics with courtesy to CPU use and other resource use information. See Document 243755.1 Implementing and Using the PL/SQL Profiler.
For serve benefit see:
Hanganalyze - These should be sent to Oracle Support Services to interpret.
Errorstacks - These should be sent to Oracle Support Services to appreciate (Some of the calls on the smoke-stack are broad and as a outcome of how an errorstack functions so , if searched for on Metalink, can lead to improper analysis.
Document 62143.1 Understanding and Tuning the Shared Pool
Document 296377.1 Handling and solution unshared cursors/large version_counts
See Document 34592.1 WAITEVENT: "log record sync" Reference Note.
It is endorsed that customers experiencing log record sync problems on 10.2.0.3 proactively request the vegetable patch for Bug 5896963 .
Typical solutions are:-
Document 34405.1 WAITEVENT: "buffer active waits" Reference Note
Document 42152.1 LATCH: CACHE BUFFERS CHAINS
Document 155971.1 Ext/Pub Resolving Intense and "Random" Buffer Busy Wait Performance Problems:
Document 163424.1 Ext/Pub How To Identify a Hot Block Within The Database Buffer Cache. TX - Document 62354.1 TX Transaction thatch - Example wait for scenarios TM - Document 33453.1 REFERENTIAL INTEGRITY AND LOCKING
Typically this is caused by two (or more) unsuitable operations being run simltaneously. Document 278316.1 Potential reasons for "WAITED TOO LONG FOR A ROW CACHE ENQUEUE LOCK! "
The reason 'enqueue hash bondage bolt waits' are here is that, typically, during deadlock showing (ie the routine Oracle uses to establish if a deadlock obviously exists), there is a complicated need is to bolt that can result in problems for other sessions. If there is a complaint with this latch, examine if a keep track record is generated is to ORA-60 and finish that issue.
No comments:
Post a Comment