[Gluster-devel] Cannot run VMware Virtual Machines on GlusterFS

[Tomoaki Sato]

Avati,

>   What is /baz/ which is prefixed behind foo as /baz/foo in the logs, but does not show up in your ls -lR?

'bar' is a gluster volume.
'baz' is a directory on the volume.
The ESXi host mounts 192.168.1.136:/bar/baz on /vmfs/volumes/1668aab4-a78e64bd directory.

Existence of 'baz' directory is not required for reproduce the issue.
Would you like to see log files without 'baz' directory too?

>Can you also send me ls -liR so that I can match inode numbers from the trace?

on the ESXi host:

~ # ls -liR /vmfs/volumes/1668aab4-a78e64bd
/vmfs/volumes/1668aab4-a78e64bd:
-1374929331 drwxr-xr-x    1 root     root               4096 Jun 19 06:41 foo

/vmfs/volumes/1668aab4-a78e64bd/foo:
~ #

on a Linux NFS client(CentOS 5.6 x86_64):

# ls -liR /vmfs/volumes/1668aab4-a78e64bd
/vmfs/volumes/1668aab4-a78e64bd:
total 8
2920037965 drwxr-xr-x 2 root root 4096 Jun 19 15:41 foo

/vmfs/volumes/1668aab4-a78e64bd/foo:
total 0
#

Regards,

Tomo

(2012/06/20 15:34), Anand Avati wrote:
> Tomoaki,
>   What is /baz/ which is prefixed behind foo as /baz/foo in the logs, but does not show up in your ls -lR? Can you also send me ls -liR so that I can match inode numbers from the trace?
>
> Avati
>
> On Tue, Jun 19, 2012 at 9:28 PM, Tomoaki Sato <tsato at valinux.co.jp <mailto:tsato at valinux.co.jp>> wrote:
>
>     Avati,
>
>
>           I saw your log and there is nothing suspicious in it. Are you sure the log captures the event where 'getcwd' returned a failure?
>
>
>     Yes, I am.
>
>
>         Can you also return the output of 'ls -lR of the data set' (from a dir level above) along with the getcwd and capture trace nfs logs from the full session? [and keep the trace nfs logs restricted to just these commands].
>
>
>     please find attached.
>
>     Regards,
>
>     Tomo
>
>
>     (2012/06/20 13:17), Anand Avati wrote:
>
>         Tomosaki,
>           I saw your log and there is nothing suspicious in it. Are you sure the log captures the event where 'getcwd' returned a failure? Can you also return the output of 'ls -lR of the data set' (from a dir level above) along with the getcwd and capture trace nfs logs from the full session? [and keep the trace nfs logs restricted to just these commands].
>
>         Thanks!
>         Avati
>
>         On Tue, Jun 19, 2012 at 7:20 PM, Tomoaki Sato <tsato at valinux.co.jp <mailto:tsato at valinux.co.jp> <mailto:tsato at valinux.co.jp <mailto:tsato at valinux.co.jp>>> wrote:
>
>             Vijay,
>
>             Thank you for your attention to this matter.
>
>             I could not find strace command on the ESXi hosts.
>             CentOS version of strace command says "strace+ ptrace(PTRACE_TRACEME, ...): Function not implemented" on the ESXi host.
>
>             Please find attached log file.
>             This log file is smaller than the previous version and covers only the 'getcwd' period.
>
>             Regards,
>
>             Tomo
>
>
>             (2012/06/20 10:51), Vijay Bellur wrote:
>
>                 On 06/18/2012 11:55 PM, Tomoaki Sato wrote:
>
>                     Vijay,
>
>                     please find attached test program and log file.
>
>                     test-code:
>
>                     $ cat getcwd.c
>                     #include <stdio.h>
>                     #include <unistd.h>
>
>                     int main(int argc, char* argv[])
>                     {
>                     char cwd[1024];
>
>                     if ( argc != 2 ) {
>                     printf("usage: %s dir\n", argv[0]);
>                     _exit(1);
>                     }
>
>                     if ( chdir(argv[1]) != 0 ) {
>                     perror("chdir");
>                     _exit(1);
>                     }
>                     if ( getcwd(cwd, 1024) == NULL ) {
>                     perror("getcwd");
>                     _exit(1);
>                     }
>                     printf("%s\n", getcwd(cwd, 1024));
>
>                     _exit(0);
>                     }
>
>                     operations:
>
>                     ~ # ./getcwd /vmfs/volumes/1668aab4-____a78e64bd/foo
>
>                     getcwd: No such file or directory
>                     ~ #
>
>
>                 Thanks for providing this test program. Can you please provide the output of strace -f ./getcwd /vmfs/volumes/1668aab4-____a78e64bd/foo and the corresponding nfs log file?
>
>
>                 -Vijay
>
>
>
>             _________________________________________________
>             Gluster-devel mailing list
>         Gluster-devel at nongnu.org <mailto:Gluster-devel at nongnu.org> <mailto:Gluster-devel at nongnu.__org <mailto:Gluster-devel at nongnu.org>>
>         https://lists.nongnu.org/__mailman/listinfo/gluster-devel <https://lists.nongnu.org/mailman/listinfo/gluster-devel>
>
>
>
>