Postgres Pro
Downloads
Home   >   mailing lists

Doubts after evaluating Xlogdump - Mailing list pgsql-general

From Vinay Gupta
Subject Doubts after evaluating Xlogdump
Date
Msg-id CAEzDN6yFcoEYSwMurBpnsod1QXcqsjQCD9uzP4r7d8Bo3mpUSg@mail.gmail.com
Whole thread Raw
Responses Re: Doubts after evaluating Xlogdump  (Jeff Janes <jeff.janes@gmail.com>)
List pgsql-general
Tree view
Hi,

Is dere any better way we can understand the contents of pg_xlog files. I am using xlogdump for analysing my xlog contents. Sometimes in my database server in some particular hours, the pgxlog files generation exceeded than the expected counts almost 3X than the  normal count. 

--- Size of 1 hour xlog files (620 files)
 
[root@db xlogdump1]# du -sh /tmp/pgxlogdump/
9.9G    /tmp/pgxlogdump/
[root@db1002 xlogdump1]# 

--- Parsing Xlog files

[root@db1002 xlogdump1]# /usr/pgsql-9.2/bin/xlogdump  /tmp/pgxlogdump/* > xlog_analysis24feb2014_18.txt
Unable to read continuation page?
Unable to read continuation page?
...
..  so on

-- Final output file size after parsing

[root@db xlogdump1]# du -sh xlog_analysis24feb2014_18.txt
3.6G    xlog_analysis24feb2014_18.txt

I don't think it parsed all files completely because output file size is too small.

So to analyse the output file , i use the below command :

cat xlog_analysis24feb2014_18.txt |awk  '{print $8 "," $7}'  | sort | uniq -dc | sort


that gives me output like below :-
 195068 s/d/r:1663/1228184/1363132,clean:
 200548 s/d/r:1663/1228184/1363155,bkpblock[1]:
 203964 s/d/r:1663/112743/5263051,insert:
 205003 s/d/r:1663/112743/5263054,insert_leaf:
 205673 s/d/r:1663/112743/5263053,insert_leaf:
 208101 s/d/r:1663/112743/5263208,insert:
 269840 ,vacuum:
 210556 s/d/r:1663/112743/5263212,insert_leaf:
 250651 s/d/r:1663/1228184/1522128,update:
 250867 s/d/r:1663/1228184/1522132,insert_leaf:
 253874 s/d/r:1663/1228184/1291583,insert_leaf:
 253887 s/d/r:1663/1228184/1291579,update:
 280151 s/d/r:1663/112743/919064,insert_leaf:
 281578 s/d/r:1663/112743/919027,update:
 293516 s/d/r:1663/1228184/1363132,bkpblock[1]:
 351713 s/d/r:1663/1228184/1363648,hot_update:
 361884 s/d/r:1663/1228184/5263067,insert_leaf:
 382961 s/d/r:1663/1228184/1363654,bkpblock[1]:
 392147 s/d/r:1663/1228184/1363652,bkpblock[1]:
 392957 s/d/r:1663/1228184/5263063,insert:
 393086 s/d/r:1663/1228184/1522150,update:
 412499 s/d/r:1663/1228184/5263286,insert_leaf:
 437279 s/d/r:1663/1228184/5263282,insert:
 927606 s/d/r:1663/1228184/1363648,update:
 932525 s/d/r:1663/1228184/1363652,insert_leaf:
 932668 s/d/r:1663/1228184/1363654,insert_leaf:
1078769 s/d/r:1663/1228184/5263108,insert_leaf:
1155642 s/d/r:1663/1228184/5263105,insert:
1181241 s/d/r:1663/1228184/5263107,insert_leaf:

Is dere any better way where i can analyse who is contributing most in my pgxlog file generation.

Also,  how can i link my insert_leaf statements to my tables/indexes ?

1078769 s/d/r:1663/1228184/5263108,insert_leaf:



Thanks

pgsql-general by date:

Previous
From: Adrian Klaver
Date:
Subject: Re: Issue with default values and Rule
Next
From: Scott Marlowe
Date:
Subject: Re: Cancelling of autovacuums considered harmful