[Gluster-users] hangs on accessing files on gluster mount (3.2.0)
mangoo at wpkg.org
Mon May 9 13:43:27 UTC 2011
On 09.05.2011 13:08, Tomasz Chmielewski wrote:
> On 09.05.2011 12:48, Mohammed Junaid Ahmed wrote:
>> Hi Tomasz,
>> Can you attach the logfiles, that will help. When does it happen - Does this happen just after the mount? What is the state of the server processes (are they busy or idle)?
> It usually happens several hours after the mount.
> Frankly, I don't see anything special in the log on either client or server.
> Only something like this relating to the "hanging" file (15064.jpeg), after I restarted gluster on the client:
> ./bricks/home-gluster-data.log-[2011-05-09 12:30:17.689988] W [socket.c:1494:__socket_proto_state_machine] 0-tcp.gluster-data-server: reading from socket failed. Error (Transport endpoint is not connected), peer (192.168.1.30:1021)
> ./bricks/home-gluster-data.log:[2011-05-09 12:30:17.690107] I [server-helpers.c:485:do_fd_cleanup] 0-gluster-data-server: fd cleanup on /www/storage/article/2307/15064.jpeg
> ./bricks/home-gluster-data.log:[2011-05-09 12:30:17.690154] I [server-helpers.c:485:do_fd_cleanup] 0-gluster-data-server: fd cleanup on /www/storage/article/2307/15064.jpeg
> ./bricks/home-gluster-data.log-[2011-05-09 12:30:17.690180] I [server-helpers.c:485:do_fd_cleanup] 0-gluster-data-server: fd cleanup on /www/storage_cache/cache/photos/link/21/21_6.jpg
> ./bricks/home-gluster-data.log-[2011-05-09 12:30:17.690203] I [server.c:438:server_rpc_notify] 0-gluster-data-server: disconnected connection from 192.168.1.30:1021
> ./bricks/home-gluster-data.log-[2011-05-09 12:30:17.690235] I [server-helpers.c:783:server_connection_destroy] 0-gluster-data-server: destroyed connection of web1-4803-2011/05/06-23:25:46:900589-gluster-data-client-0
> ./bricks/home-gluster-data.log-[2011-05-09 12:30:31.589199] I [server-handshake.c:534:server_setvolume] 0-gluster-data-server: accepted client from 192.168.1.30:1017
> If it matters, I was copying hundreds of thousand of files from a different gluster client in the meantime, using rsync
> (which creates a temporary file and renames it upon completion) - not sure if it can be related.
> I stopped copying now to see if the issue still shows up.
It hangs also when I don't use rsync.
It's something fishy here.
Are such stalls while reading files known with glusterfs 3.2.0?
More information about the Gluster-users