Xlogdump compiling error : undefined reference to `ber_sockbuf_io_udp'

On Sun, Feb 9, 2014 at 10:02 PM, Adarsh Sharma <eddy.adarsh@gmail.com> wrote:

I need to use xlogdump to analyse my xlog files in postgresql to understand
why there are so much xlog generation on Master.

Only PostgreSQL is supported. EDB's product is not.

--
Regards,
Peter Geoghegan

--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general

Hi Peter,

Thanks for your reply. I Complied xlogdump successfully with Pg-9.2 and now
i am using it to parse EDB's pg_xlog files and i am able to do that. Below
are the details :

[cur:1835/26FAE218, xid:471303469, rmid:9(Heap2), len/tot_len:24/6900,
info:24, prev:1835/26FAD6D8] bkpblock[1]:
s/d/r:1663/1228184/*1363132*blk:58795 hole_off/len:396/1372
[cur:1835/26FAFD10, xid:471303469, rmid:10(Heap), len/tot_len:26/58,
info:96, prev:1835/26FAE218] lock exclusive: s/d/r:1663/1228184/1363132
block 58795 off 48
[cur:1835/26FAFD50, xid:471303469, rmid:10(Heap), len/tot_len:314/346,
info:32, prev:1835/26FAFD10] update: s/d/r:1663/1228184/1363132 block 58795
off 48 to block 58795 off 49
[cur:1835/26FAFEB0, xid:471303469, rmid:11(Btree), len/tot_len:34/66,
info:0, prev:1835/26FAFD50] insert_leaf: s/d/r:1663/1228184/1363154 tid
13880/179
[page:2008, xlp_info:5, xlp_tli:1, xlp_pageaddr:1835/26FB0000]
XLP_FIRST_IS_CONTRECORD XLP_BKP_REMOVABLE
[cur:1835/26FAFEF8, xid:471303469, rmid:11(Btree), len/tot_len:18/5754,
info:8, prev:1835/26FAFEB0] insert_leaf: s/d/r:1663/1228184/1363155 tid
28366/20
[cur:1835/26FB1988, xid:471303469, rmid:11(Btree), len/tot_len:34/66,
info:0, prev:1835/26FB15C8] insert_leaf: s/d/r:1663/1228184/1363154 tid
13880/180
[cur:1835/26FEBD58, xid:471303469, rmid:11(Btree), len/tot_len:50/82,
info:0, prev:1835/26FEB9B0] insert_leaf: s/d/r:1663/1228184/*5144337* tid
3491/103
[cur:1835/26FEBDB0, xid:471303469, rmid:11(Btree), len/tot_len:34/66,
info:0, prev:1835/26FEBD58] insert_leaf: s/d/r:1663/1228184/*5144339* tid
76/328
[cur:1835/26FB19D0, xid:471303469, rmid:11(Btree), len/tot_len:18/5070,
info:8, prev:1835/26FB1988] bkpblock[1]: s/d/r:1663/1228184/1363155
blk:38821 hole_off/len:108/3196
[page:2010, xlp_info:5, xlp_tli:1, xlp_pageaddr:1835/26FB4000]
XLP_FIRST_IS_CONTRECORD XLP_BKP_REMOVABLE
[cur:1835/26FB2DB8, xid:471303469, rmid:9(Heap2), len/tot_len:24/6704,
info:24, prev:1835/26FB19D0] clean: s/d/r:1663/1228184/1363132 block:49662
redirected/dead/unused:0/2/0 removed xid:471257627

Is dere any document where i can co-relate this tid with relations/indexes
in the database. I tried but not able to find :-

wap_prod_adserve=# SELECT relname, oid FROM pg_class where oid=1363155 ;
relname | oid
---------+-----
(0 rows)

wap_prod_adserve=# SELECT relname, oid FROM pg_class where oid=1363154;
relname | oid
---------+-----
(0 rows)

wap_prod_adserve=# SELECT relname, oid FROM pg_class where oid=5144339;
relname | oid
---------+-----
(0 rows)

I just want to know the reason of too much xlog file generation.

Thanks

Below link is little helpful :-

http://michael.otacoo.com/postgresql-2/postgres-9-3-feature-highlight-pg_xlogdump/

Thanks

On Wed, Feb 12, 2014 at 5:19 PM, Adarsh Sharma <eddy.adarsh@gmail.com> wrote:

Below link is little helpful :-

http://michael.otacoo.com/postgresql-2/postgres-9-3-feature-highlight-pg_xlogdump/

Postgres core includes pg_xlogdump since 9.3 and not xlogdump. As
their outputs are a bit different you should directly contact the
maintainer xlogdump if you begin playing with it. Saying that, I'd
guess that the output of s/d/r means tablespace/database/relation. So
if you cannot find those oids it means that they have been deleted at
a point later than the record you are looking at.

Regards,
--
Michael

--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general

Adarsh Sharma escribi�:

[cur:1835/26FAFEF8, xid:471303469, rmid:11(Btree), len/tot_len:18/5754, info:8, prev:1835/26FAFEB0] insert_leaf: s/d/r:1663/1228184/1363155 tid 28366/20

Is dere any document where i can co-relate this tid with relations/indexes
in the database. I tried but not able to find :-

wap_prod_adserve=# SELECT relname, oid FROM pg_class where oid=1363155 ;
relname | oid
---------+-----
(0 rows)

Probably the reason you don't find any rows is because you're querying
the wrong column. The "r" stands for relfilenode, and thus you must use
relfilenode in the WHERE clause, not oid:

SELECT relname, oid FROM pg_class where relfilenode=1363155 ;

--
�lvaro Herrera http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general

Yes. Able to retrieve table names now.

I am reviewing the xlogdump output. Is dere any way where i can understand
what is the meaning of these keywords in it.

[root@1002 tmp]# xlogdump pg_xlog/* > fullanalysis.txt
[root@1002 tmp]# cat fullanaysis.txt | awk '{print $7}' | sort | uniq
XLP_BKP_REMOVABLE
bkpblock[1]:
bkpblock[2]:
commit_compact
create
delete:
heap2_multi_insert:
hot_update:
inplace:
insert_leaf:
insert_upper:
newpage:
newroot:
offset
reuse_page:
seq
split_l:
split_l_root:
split_r:
split_r_root:
standby
update(init):
update:

Thanks

On Fri, Feb 14, 2014 at 3:14 PM, Adarsh Sharma <eddy.adarsh@gmail.com> wrote:

Yes. Able to retrieve table names now.

I am reviewing the xlogdump output. Is dere any way where i can understand
what is the meaning of these keywords in it.

[root@1002 tmp]# xlogdump pg_xlog/* > fullanalysis.txt
[root@1002 tmp]# cat fullanaysis.txt | awk '{print $7}' | sort | uniq
XLP_BKP_REMOVABLE
bkpblock[1]:
bkpblock[2]:
commit_compact
create
delete:
heap2_multi_insert:
hot_update:
inplace:
insert_leaf:
insert_upper:
newpage:
newroot:
offset
reuse_page:
seq
split_l:
split_l_root:
split_r:
split_r_root:
standby
update(init):
update:

Even if it is easy to guess to which WAL operations all those terms
refer to, referring to the source code is easier IMO:
https://github.com/snaga/xlogdump/blob/master/xlogdump_rmgr.c
Regards,
--
Michael

--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general