[HACKERS] On markers of changed data

Hi, hackers!

Currently I'm working on page-level incremental backups using WAL-G codebase[0]. And I have two questions that I cannot
resolvemyself. 

Incremental backup is a set of changes, that should be applied over preexisting backup. I use page LSN to understand
shouldpage be backup`ed or not. 

Question 1. FSM and VM.
As you can see here [1] FSM and VM files are exempt from incremental tracking and are backuped as whole files. I've
doneit this way, because sanity checks [2] of page headers have indicated a lot of "invalid" pages in FSM and VM files.
Butseems like in some pages headers are valid with sane LSNs. 
Can I use LSNs as history marker on FSM and VM pages? On 1Tb backup I get like 150Mb of FSM+VM, and it's kind of a lot.

Question 2. File dates.
Is it safe to use file modification time to track that file were changes since previous backup? If the file has date
beforestart of previous backup I just add it to "skip list" [3]. 
I have assumption: every time file is changes in filesystem, it's modification date is updated to higher value.
Is this assumption valid for most of used platforms and filesystems? Or can I check this "capacity" of FS?

Thank you for your attention. I'll be glad to receive any information\pointers on this matter.


Best regards, Andrey Borodin, Yandex.

[0] https://github.com/wal-g/wal-g/pull/29
[1] https://github.com/wal-g/wal-g/pull/29/files#diff-d77406e827f5f947d4d4a1e6d76c1f4eR114
[2] https://github.com/wal-g/wal-g/pull/29/files#diff-d77406e827f5f947d4d4a1e6d76c1f4eR50
[3] https://github.com/wal-g/wal-g/pull/29/files#diff-f5c8f0067297f98eb5acc6e2c6b1b234R87

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