The invalid argument on chmod in fuse is fixed here:

http://www.orangefs.org/trac/orangefs/changeset/9212

Patch for that change:

$ svn diff -r 9211:9212 http://www.orangefs.org/svn/orangefs/branches/stable
Index: src/apps/fuse/pvfs2fuse.c
===================================================================
--- src/apps/fuse/pvfs2fuse.c   (revision 9211)
+++ src/apps/fuse/pvfs2fuse.c   (revision 9212)
@@ -507,8 +507,10 @@
    ret = lookup( path, &pfh, PVFS2_LOOKUP_LINK_FOLLOW );
    if ( ret < 0 )
          return PVFS_ERROR_TO_ERRNO_N( ret );
-   
-   new_attr.perms = mode;
+   /* FUSE passes in 5 octets in 'mode'. However, the the first 
+    * octet is not related to permissions, hence checking only
+    *  the lower 4 octets */
+   new_attr.perms = mode & 07777;
    new_attr.mask = PVFS_ATTR_SYS_PERM;
  
    ret = PVFS_sys_setattr(pfh.ref,new_attr,&pfh.creds);


On Mar 12, 2012, at 5:51 PM, Andrew Savchenko wrote:

> Hello,
> 
> On Mon, 12 Mar 2012 16:33:55 -0400 John Chandy wrote:
>> Jim,
>> 
>> I can provide some answers to the fuse question.  In principle, the fuse 
>> module in PVFS works the same way as the kernel module - i.e. there is still 
>> a user to kernel back to userspace transition.  The tests that we did (about 
>> 4 years ago) showed that pvfs2fuse had similar performance to the kernel 
>> module as long as the block size was less than 128K.
> 
> I tested the orangefs-2.8.5 fuse client on Linux-3.0.10 recently.
> It looks limited in functionality compared to the kernel module:
> 
> 1) chmod on any file fails:
> $ chmod 0755 test.file
> chmod: changing permissions of `test.file': Invalid argument
> 
> 2) If you multiple pvfs2-servers are running on the same node, they
> will have different config files due to different data/meta
> directories, log file and so on. The fuse client checks sizes and
> checksums of all server's configs and refuse to start if they are
> different. pvfs2-client powered by the kernel module works fine in
> this case.
> 
> I haven't performed speed tests for fuse client. Limitations above
> are sufficient for me to make a choise between two clients.
> 
> Best regards,
> Andrew Savchenko
> _______________________________________________
> Pvfs2-users mailing list
> [email protected]
> http://www.beowulf-underground.org/mailman/listinfo/pvfs2-users


_______________________________________________
Pvfs2-users mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-users

Reply via email to