Hi Matthias,
Where is the storage location located on hostX for the servers?
i.e. What's listed in your server.conf files?
You could turn on debugging on the server, to see what's happening
there. Change EventLogging in your fs.conf from "none" to "verbose",
and restart the servers.
-sam
On Jan 11, 2007, at 8:03 AM, Matthias Pfeifer wrote:
hello,
we have just set up a pvfs-2.6.1 with 8 servers host0-host7 here,
where host0 acts as a meta-data server. All hosts are linux-2.6
and have a pvfs2-server, pvfs2-client, pvfs2-client-core running
the kernel-module pvfs2 is loaded. Testing the system
with
pvfs2-cp -t 64mega.dat /mnt/pvfs2
does not terminate. tracing the pvfs2-cp process with strace
reveals that pvfs2-cp is doing
epoll_wait(4, {}, 16, 10) = 0
(this notation is output from strace) again and again.
Trying VFS-calls with
cp 64mega.dat /mnt/pvfs2
does not terminate, too. strace here shows that the cp-process
hangs in
write(4, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0
\0"..., 4194304
Might this be due to an wrong setup? What else might be wrong?
The pvfs2-logfile does not reveal anything what looks like an error.
Only these entries:
[D 01/05 12:01] PVFS2 Server version 2.6.1 starting.
matthias
_______________________________________________
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
