gluster-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gluster-devel] Segmentation fault occurred in glusterfs_getxattr.


From: Raghavendra G
Subject: Re: [Gluster-devel] Segmentation fault occurred in glusterfs_getxattr.
Date: Wed, 26 Aug 2009 14:56:11 +0400

Hi,

On Wed, Aug 26, 2009 at 1:07 PM, Minjae Shin <address@hidden> wrote:

Hello,

 

1. When I call glusterfs_getxattr, I received 'Segmentation fault' error like the following:

-------------------------------------------

Program received signal SIGSEGV, Segmentation fault.

[Switching to Thread -1209996640 (LWP 16114)]

0xb7ebc046 in free () from /lib/tls/libc.so.6

(gdb) bt

#0  0xb7ebc046 in free () from /lib/tls/libc.so.6

#1  0xb7f9c611 in __glusterfs_glh_getxattr (handle=0xb7f854c0,

    path=0x808fc80 "/r/rest_test_nodelete/hello_70301.txt",

    name=0x8090206 "hello_70301.txt", value=0x804a008, size=1024, whichop=1)

    at libglusterfsclient.c:2273

#2  0xb7f9c944 in glusterfs_getxattr (

    path=0x8048af8 "/home/minjae/gluster_test/libglusterfs/mount/r/rest_test_nodelete/hello_70301.txt", name=0x8048bbd "user.test", value=0x804a008, size=1024)

    at libglusterfsclient.c:2327

#3  0x0804886d in main () at test.c:54

-------------------------------------------

The attribute name was "user.test" in #3, but this was changed to "hello_70301.txt" in #2.

About some files, glusterfs_getxattr is working well, but sometimes working strange including this.

Progress on this bug can be tracked at
http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=228

 

 

2. This is about glusterfs_flistxattr, and here is my code :

-------------------------------------------

    ssize_t bufsize = 1024;

    char *buffer = NULL;

    buffer = calloc( bufsize, sizeof(char) );

 

    ret = glusterfs_flistxattr( fd, buffer, bufsize );

-------------------------------------------

The 'ret' is 38 with Success, but buffer is empty.

glusterfs_flistxattr is not implemented, It just returns ENOSYS. The correct way should've been returning -1 with errno set to ENOSYS. A bug is filed at
http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=229
 

 

Here is the log:

[2009-08-26 17:21:59] D [dht-common.c:161:dht_lookup_dir_cbk] distribute: fixing assignment on /

[2009-08-26 17:21:59] D [dht-selfheal.c:435:dht_selfheal_directory] distribute: 2 subvolumes down -- not fixing

[2009-08-26 17:21:59] D [libglusterfsclient.c:1410:libgf_vmp_search_entry] libglusterfsclient: VMP Entry found: /home/minjae/gluster_test/libglusterfs/mount/r/rest_test_nodelete/hello_70301.txt: /home/minjae/gluster_test/libglusterfs/mount/

[2009-08-26 17:21:59] D [libglusterfsclient-dentry.c:245:__do_path_resolve] libglusterfsclient-dentry: resolved path(/r/rest_test_nodelete/hello_70301.txt) till 1(/). sending lookup for remaining path

[2009-08-26 17:21:59] D [dht-layout.c:101:dht_layout_search] distribute: no subvolume for hash (value) = 250795985

[2009-08-26 17:21:59] D [dht-helper.c:228:dht_subvol_get_hashed] distribute: could not find subvolume for path=/r

[2009-08-26 17:21:59] D [dht-common.c:820:dht_lookup] distribute: no subvolume in layout for path=/r, checking on all the subvols to see if it is a directory

 

 

Could you check these, please?

I'm using glusterfs 2.0.6

 

Thanks in advnace!

--

Minjae Shin

 


_______________________________________________
Gluster-devel mailing list
address@hidden
http://lists.nongnu.org/mailman/listinfo/gluster-devel




--
Raghavendra G


reply via email to

[Prev in Thread] Current Thread [Next in Thread]