[Gluster-devel] [bug #19356] glusterfs-FUSE ftruncate not working

[anonymous]

URL:
  <http://savannah.nongnu.org/bugs/?19356>

                 Summary: glusterfs-FUSE ftruncate not working
                 Project: Gluster
            Submitted by: None
            Submitted on: Tuesday 03/20/2007 at 13:51 UTC
                Category: GlusterFS
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: Improper behaviour
                  Status: None
                 Privacy: Public
             Assigned to: None
             Open/Closed: Open
         Discussion Lock: Any
        Operating System: GNU/Linux

    _______________________________________________________

Details:

We have a process that writes sequentially to a file until the file overflows
the 1 GB limit that we have set.  At that point, the process does an
ftruncate() back to the end of the last entry, closes the file and opens a
new one.  

The ftruncate() call is not truncating the file.  We end up with a partial
record at the end of each file.

This works with other local filesystems and Lustre - these do not use FUSE,
of course.

System specs:
Linux 2.6.20-gentoo #3 SMP Tue Mar 6 13:19:08 CET 2007 x86_64 Dual-Core AMD
Opteron(tm) Processor 2212 HE AuthenticAMD GNU/Linux

GlusterFS v1.3.0-pre2.2 patched up to 76

FUSE info:
  6402 Feb 28 22:32 fuse-bridge.c
 62948 Feb 25 17:00 fuse-internals.c
  2440 Jan 18 04:12 fuse-internals.h
  9308 Mar  1 01:07 glusterfs.c
  1410 Feb 28 22:32 glusterfs.h

Using:
Client:
  cluster/unify
  writebehind
  ALU

Servers:
4 bricks (separate servers)
  iothreads
  posix
  




    _______________________________________________________

Reply to this item at:

  <http://savannah.nongnu.org/bugs/?19356>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.nongnu.org/