每天坐在电脑前超过4小时的朋友注意
2008-09-17 18:42
405 查看
Encourter the Potential blocker issue!
Some one said it is bug:
Bug 6495842: GES POTENTIAL BLOCKER ON RESOURCE TX CAUSE RAC TO HANG
Bug 7372971: RAC : HANG DUE TO GES DEADLOCK AND POTENTIAL BLOCKER
Generate the ADDM report and suggested to change "db_writer_processes" parameter, it was
"4" currently which cause the free buffer waits event high.
After checking the alert logs I found the Potential blocker and Mview failure SQL always
happened at same time or in the following consecutive time.
Notice the ORA-01555.
I check the unto space usage but find it only 29% full.
Then I move to check the undo retention and find it is only 900.
suspect some restart the instance with the default parameter file.
I use the follwoing SQL
SQL> select max(maxquerylen)/3600 "TIME IN HRS" from v$undostat;
TIME IN HRS
-----------
2.94944444
SQL> select max(maxquerylen) "TIME IN HRS" from v$undostat;
TIME IN HRS
-----------
10618
The biggest SQL is almost 3hours long.
So I 'd like to change the Mview refresh frequency to 1 hours to make the refresh done in shorter time and change the
undo_retention from 900 to 18000,Hope this action can avoid the error ORA-01555.
Do not find the error again after change.
Extract from the alert log.
Tue Nov 2 16:59:25 2010
GES: Potential blocker (pid=19068) on resource TX-00870007-001B3C18;
enqueue info in file
/opt/app/oracle/admin/RPTS/bdump/rpts3_lmd0_15052.trc and DIAG trace file
Tue Nov 2 16:59:25 2010
GES: Potential blocker (pid=19171) on resource FB-00000013-0C03999A;
enqueue info in file
/opt/app/oracle/admin/RPTS/udump/rpts3_ora_19068.trc and DIAG trace file
Tue Nov 2 17:18:16 2010
ORA-01555 caused by SQL statement below (SQL ID: fd958md7mtzqa, Query
Duration=6475 sec, SCN: 0x07b2.408c8ec2):
Tue Nov 2 17:18:16 2010
SELECT /*+ */
"A2"."IDENTIFIER","A2"."FIRSTOCCURRENCE","A2"."LASTOCCURRENCE","A2"."CLEAREDTIMESTAMP","A2"."INTERNALTICKETNUM","A2"."TALLY","A2"."SUMMARY","A2"."ALERTKEY","A2"."AGENT","A2"."IPADDRESS","A2"."CUSTOMERNAME","A2"."NODE","A2"."MANAGER","A2"."ALERTGROUP","A2"."GRADE","A2"."DEVSTATUS","A2"."ORIGINALMESSAGE","A2"."DELETEDAT","A2"."ORIGINALSEVERITY","A2"."SERVERNAME","A2"."SERVERSERIAL","A2"."CUSTOMERSTATUS","A2"."MANAGINGHOST","A2"."MAINTFLAG","A2"."MAST","A2"."STARTTIME","A2"."ENDTIME","A2"."CTARECEIVETIME","A2"."OWNERNAME"
FROM "REPORTER"."REPORTER_STATUS" "A2", (SELECT DISTINCT
"A3"."SERVERNAME" "SERVERNAME","A3"."SERVERSERIAL" "SERVERSERIAL" FROM
"REPORTER"."MLOG$_REPORTER_STATUS" "A3" WHERE "A3"."SNAPTIME$$">:1
AND "A3"."DMLTYPE$$"<>'D') "A1" WHERE
"A2"."SERVERNAME"="A1"."SERVERNAME" AND .......
Some one said it is bug:
Bug 6495842: GES POTENTIAL BLOCKER ON RESOURCE TX CAUSE RAC TO HANG
Bug 7372971: RAC : HANG DUE TO GES DEADLOCK AND POTENTIAL BLOCKER
Generate the ADDM report and suggested to change "db_writer_processes" parameter, it was
"4" currently which cause the free buffer waits event high.
After checking the alert logs I found the Potential blocker and Mview failure SQL always
happened at same time or in the following consecutive time.
Notice the ORA-01555.
I check the unto space usage but find it only 29% full.
Then I move to check the undo retention and find it is only 900.
suspect some restart the instance with the default parameter file.
I use the follwoing SQL
SQL> select max(maxquerylen)/3600 "TIME IN HRS" from v$undostat;
TIME IN HRS
-----------
2.94944444
SQL> select max(maxquerylen) "TIME IN HRS" from v$undostat;
TIME IN HRS
-----------
10618
The biggest SQL is almost 3hours long.
So I 'd like to change the Mview refresh frequency to 1 hours to make the refresh done in shorter time and change the
undo_retention from 900 to 18000,Hope this action can avoid the error ORA-01555.
Do not find the error again after change.
Extract from the alert log.
Tue Nov 2 16:59:25 2010
GES: Potential blocker (pid=19068) on resource TX-00870007-001B3C18;
enqueue info in file
/opt/app/oracle/admin/RPTS/bdump/rpts3_lmd0_15052.trc and DIAG trace file
Tue Nov 2 16:59:25 2010
GES: Potential blocker (pid=19171) on resource FB-00000013-0C03999A;
enqueue info in file
/opt/app/oracle/admin/RPTS/udump/rpts3_ora_19068.trc and DIAG trace file
Tue Nov 2 17:18:16 2010
ORA-01555 caused by SQL statement below (SQL ID: fd958md7mtzqa, Query
Duration=6475 sec, SCN: 0x07b2.408c8ec2):
Tue Nov 2 17:18:16 2010
SELECT /*+ */
"A2"."IDENTIFIER","A2"."FIRSTOCCURRENCE","A2"."LASTOCCURRENCE","A2"."CLEAREDTIMESTAMP","A2"."INTERNALTICKETNUM","A2"."TALLY","A2"."SUMMARY","A2"."ALERTKEY","A2"."AGENT","A2"."IPADDRESS","A2"."CUSTOMERNAME","A2"."NODE","A2"."MANAGER","A2"."ALERTGROUP","A2"."GRADE","A2"."DEVSTATUS","A2"."ORIGINALMESSAGE","A2"."DELETEDAT","A2"."ORIGINALSEVERITY","A2"."SERVERNAME","A2"."SERVERSERIAL","A2"."CUSTOMERSTATUS","A2"."MANAGINGHOST","A2"."MAINTFLAG","A2"."MAST","A2"."STARTTIME","A2"."ENDTIME","A2"."CTARECEIVETIME","A2"."OWNERNAME"
FROM "REPORTER"."REPORTER_STATUS" "A2", (SELECT DISTINCT
"A3"."SERVERNAME" "SERVERNAME","A3"."SERVERSERIAL" "SERVERSERIAL" FROM
"REPORTER"."MLOG$_REPORTER_STATUS" "A3" WHERE "A3"."SNAPTIME$$">:1
AND "A3"."DMLTYPE$$"<>'D') "A1" WHERE
"A2"."SERVERNAME"="A1"."SERVERNAME" AND .......
相关文章推荐
- 每天坐在电脑前超过4小时的朋友注意
- 送给每天用电脑超4小时的朋友需要注意了
- [注意]如果你每天电脑前超过4小时 应当这样 [图组]
- 给每天在电脑面前超过4小时的朋友
- 给每天在电脑面前超过4小时的朋友 zz
- 如果你一天有超过4小时都在电脑前,那么要注意了!请跟着做!
- 如果你每天面对电脑超过4小时应该这样
- 每天电脑前超过四小时者要注意哦 ZT
- 我的第100篇Post:〔转〕给每天在电脑前4个小时以上的朋友
- 送给每天用电脑超4小时的朋友
- 如果你每天电脑前超过4小时 应当这样
- 最近戴着眼镜坐电脑前总是不自觉的眼痛就搜了下怎么保护眼睛无意中看到了这篇文章希望广大爱好编程的朋友多注意保护自己的眼睛!!
- 送给每天用电脑超4小时的朋友!!
- 送给每天用电脑超4小时的朋友
- 送给每天用电脑超4小时的朋友
- 送给每天用电脑超4小时的朋友
- 电脑前的朋友注意坐姿 --详图解说
- 送给每天用电脑超过4小时的朋友们!
- 为了眼睛的健康改变电脑的色调-献给长期坐在电脑前的朋友
- 用DELL笔记本电脑的朋友们注意了!你们随时有可能被怀疑是恐怖分子,而被捉进监狱!...