Re: [REVIEW] Re: Compression of full-page-writes

>I think this was changed based on following, if I am not wrong. 

>http://www.postgresql.org/message-id/54297A45.8080904@...
Yes this change is the result of the above complaint.

>Attaching the compression status to XLogRecord is more 
>in-line with the fact that all the blocks are compressed, and not each 
>one individually, so we basically now duplicate an identical flag 
>value in all the backup block headers, which is a waste IMO. 
>Thoughts? 

If I understand your point correctly, as all blocks are compressed, adding
compression attribute to XLogRecord surely makes more sense if the record
contains backup blocks . But in case of XLOG records without backup blocks
the compression attribute in record header might not make much sense.

Attaching the status of compression to XLogRecord will mean that the status
is duplicated across all records. It will mean that it is an attribute of
all the records when it is only an attribute of records with backup blocks
or the attribute of backup blocks. 
The current approach is adopted with this thought.


Regards,
Rahila Syed



--
View this message in context: http://postgresql.nabble.com/Compression-of-full-page-writes-tp5769039p5826487.html
Sent from the PostgreSQL - hackers mailing list archive at Nabble.com.