Sync vs. fsync during checkpoint - Mailing list pgsql-hackers

From Bruce Momjian
Subject Sync vs. fsync during checkpoint
Date
Msg-id 200401301701.i0UH1bZ08600@candle.pha.pa.us
Whole thread Raw
Responses Re: [pgsql-hackers-win32] Sync vs. fsync during checkpoint  (Tom Lane <tgl@sss.pgh.pa.us>)
Re: Sync vs. fsync during checkpoint  (Greg Stark <gsstark@mit.edu>)
List pgsql-hackers
As some know, win32 doesn't have sync, and some are concerned that sync
isn't reliable enough during checkpoint anyway.

The trick is to somehow record all files modified since the last
checkpoint, and open/fsync/close each one.   My idea is to stat() each
file in each directory and compare the modify time to determine if the
file has been modified since the last checkpoint.  I can't think of an
easier way to efficiently collect all modified files.  In this case, we
let the file system keep track of it for us.

However, on XP, I just tested if files that are kept open have their
modification times modified, and it seems they don't.  If I do:

    while :
        echo test
        sleep 5
    done > x

I see the file size grow every 5 seconds, but I don't see the
modification time change.  Can someone confirm this?

--
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

pgsql-hackers by date:

Previous
From: "BARTKO, Zoltan"
Date:
Subject: Re: msg translation into sk_SK, Docs: SGML -> XML
Next
From: Scott Lamb
Date:
Subject: Re: Mixing threaded and non-threaded