[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gluster-bugs
Subject:    [Bugs] [Bug 1773476] gluster does not return correct filesize and blocksize after ftruncate
From:       bugzilla () redhat ! com
Date:       2019-11-29 4:55:09
Message-ID: bug-1773476-371520-y7BTZs4Bv2 () https ! bugzilla ! redhat ! com/
[Download RAW message or body]

https://bugzilla.redhat.com/show_bug.cgi?id=1773476



--- Comment #12 from Ravishankar N <ravishankar@redhat.com> ---
(In reply to Raghavendra G from comment #10)
(In reply to Raghavendra G from comment #11)

> Just to get more details, when was the last stat (with all fields valid) sent?
> Is the 1s you mentioned was,
> 1. after the lookup/readdirp response that returned zeroed out stat (except
> for gfid and type)?
> 2. after a valid lookup/readdirp/stat/fstat response that returned a
> completely valid stat (all fields are valid)?

For the test "[root@vm1 fuse_mnt]# truncate -s 52 $FILE && stat -c "%s" $FILE",
here are the fops and their counts from profile info:
Note: there is no readdir(p) involved.
With readdir-ahead on:
------------------------
1     RELEASE
1       FLUSH
2    FINODELK
1        OPEN
1   FTRUNCATE
2    FXATTROP
2      LOOKUP

With readdir-ahead disabled:
--------------------------

 1     RELEASE
 2    FINODELK
 1       FLUSH
 1        OPEN
 1   FTRUNCATE
 2    FXATTROP
 1    GETXATTR
 4      LOOKUP    

So with readdir-ahead disabled (i.e. rda_inode_ctx_update_iatts() is not
involved), there are 2 extra lookups after the ftruncate that helps getting the
correct size. 
With readdir ahead enabled, those 2 lookup do not seem to be coming which means
the kernel is still considering the values returned during the truncate cbk as
valid.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.
_______________________________________________
Bugs mailing list
Bugs@gluster.org
https://lists.gluster.org/mailman/listinfo/bugs
[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic