From: | Harry Yau <harry(at)regaltronic(dot)com> |
---|---|
To: | pgsql-admin(at)postgresql(dot)org |
Subject: | WAL Question |
Date: | 2001-07-04 01:37:22 |
Message-ID: | 3B427351.5581945D@regaltronic.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
Hi Dear:
I am using Postgres 7.0 on RedHat Linux 6.3 right now. I wanna know
is there transaction log in PostGres 7.0? I know there is WAL in 7.1. I
am also thinking to upgrade to 7.1.2 too. I have read a lot of message
in the PostgreSQL mailing list archive which is about the WAL.
The following is some conclusion of what I learn from there:
* checkpoint is made according to the config setting of
checkpoint_segment and checkpoint_timeout.
* after a checkpoint has been made, the log flush to the disk drive.
* after a checkpoint has been made, any log segments writtend before
the redo record ( the record in the log whick the REDO operation
begin) are REMOVE to free disk space in WAL directory.
* seems to me, everything is automatic.
Please, point me out if there anything wrong.
However, there is still something really confusing me. In PostgreSQL
7.1 Documentation Chapter 9 WAL , 9.2.1 Database Recovery with WAL, it
tell me about the flow of the process to recover from WAL: 1. reads
pg_control, 2. reads checkpoint record, 3. reads redo record, and then
finally 4. REDO operation. I wonder do we need to run all the above
procedures manually, after a crash?? or They will REALLY be
automatically run when re-start the postgres after a crash?? It sounds
like a miracle for me!!
Thank You very Much
Harry Yau
From | Date | Subject | |
---|---|---|---|
Next Message | Sudheer Palapparambil | 2001-07-04 07:46:20 | RUNNING STANDALONE |
Previous Message | Jie Liang | 2001-07-03 21:27:54 | Re: Cannot drop database that is in use (NOT) |