Bug 803663

Summary: Sever failed to create reply frame because of connection object being NULL
Product: [Community] GlusterFS Reporter: Shwetha Panduranga <shwetha.h.panduranga>
Component: protocolAssignee: Amar Tumballi <amarts>
Status: CLOSED WONTFIX QA Contact:
Severity: high Docs Contact:
Priority: high    
Version: mainlineCC: gluster-bugs, vraman
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2012-04-17 18:11:51 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
Brick Log File none

Description Shwetha Panduranga 2012-03-15 11:44:07 UTC
Created attachment 570265 [details]
Brick Log File

Description of problem:

Server failed to create frame because of connection object being NULL and doesn't send reply back to client . Hence client application hangs waiting for a response from server. 

[2012-03-14 23:33:43.510756] E [server-helpers.c:735:server_alloc_frame] (-->/usr/local/lib/libgfrpc.so.0(rpcsvc_handle_rpc_call+0x360) [0x7f679d2970a9] (-->/usr/local/
lib/glusterfs/3.3.0qa27/xlator/protocol/server.so(server_finodelk+0xcb) [0x7f6793bd53db] (-->/usr/local/lib/glusterfs/3.3.0qa27/xlator/protocol/server.so(get_frame_from
_request+0x89) [0x7f6793bc11c5]))) 0-server: invalid argument: conn
[2012-03-14 23:33:43.510836] W [rpcsvc.c:1111:rpcsvc_error_reply] (-->/usr/local/lib/libgfrpc.so.0(rpc_transport_notify+0x130) [0x7f679d29cda8] (-->/usr/local/lib/libgf
rpc.so.0(rpcsvc_notify+0x181) [0x7f679d29744c] (-->/usr/local/lib/libgfrpc.so.0(rpcsvc_handle_rpc_call+0x37b) [0x7f679d2970c4]))) 0-: sending a RPC error reply
[2012-03-14 23:33:43.511039] E [server-helpers.c:735:server_alloc_frame] (-->/usr/local/lib/libgfrpc.so.0(rpcsvc_handle_rpc_call+0x360) [0x7f679d2970a9] (-->/usr/local/
lib/glusterfs/3.3.0qa27/xlator/protocol/server.so(server_finodelk+0xcb) [0x7f6793bd53db] (-->/usr/local/lib/glusterfs/3.3.0qa27/xlator/protocol/server.so(get_frame_from
_request+0x89) [0x7f6793bc11c5]))) 0-server: invalid argument: conn
[2012-03-14 23:33:43.511129] W [rpcsvc.c:1111:rpcsvc_error_reply] (-->/usr/local/lib/libgfrpc.so.0(rpc_transport_notify+0x130) [0x7f679d29cda8] (-->/usr/local/lib/libgf
rpc.so.0(rpcsvc_notify+0x181) [0x7f679d29744c] (-->/usr/local/lib/libgfrpc.so.0(rpcsvc_handle_rpc_call+0x37b) [0x7f679d2970c4]))) 0-: sending a RPC error reply
[2012-03-14 23:33:43.511596] I [socket.c:2377:socket_submit_reply] 0-tcp.dstore1-server: not connected (priv->connected = 255)
[2012-03-14 23:33:43.511648] E [rpcsvc.c:1078:rpcsvc_submit_generic] 0-rpc-service: failed to submit message (XID: 0x233872x, Program: GlusterFS 3.3.0qa27, ProgVers: 330, Proc: 30) to rpc-transport (tcp.dstore1-server)
[2012-03-14 23:33:43.512071] I [socket.c:2377:socket_submit_reply] 0-tcp.dstore1-server: not connected (priv->connected = 255)
[2012-03-14 23:33:43.512104] E [rpcsvc.c:1078:rpcsvc_submit_generic] 0-rpc-service: failed to submit message (XID: 0x23261x, Program: GlusterFS 3.3.0qa27, ProgVers: 330, Proc: 27) to rpc-transport (tcp.dstore1-server)
[2012-03-14 23:33:43.518857] E [server.c:166:server_submit_reply] (-->/usr/local/lib/libglusterfs.so.0(default_finodelk_cbk+0x14d) [0x7f679d4c97ca] (-->/usr/local/lib/glusterfs/3.3.0qa27/xlator/debug/io-stats.so(io_stats_finodelk_cbk+0x23a) [0x7f6793df2f2b] (-->/usr/local/lib/glusterfs/3.3.0qa27/xlator/protocol/server.so(server_finodelk_cbk+0x25e) [0x7f6793bc4f6a]))) 0-: Reply submission failed


Version-Release number of selected component (if applicable):
3.3.0qa27

How reproducible:
often

Steps to Reproduce:
1.create a replicate volume (1 X 3)
2.create fuse mounts from client
3.run ping_pong test from fuse mount: ping_pong f13 -rw 100 50 300
4.bring down a brick while ping_pong is running
5.bring back the brick (changes the graph)

Actual results:
ping_pong test hangs on fuse mount when a brick goes down and come back online.

Comment 1 Amar Tumballi 2012-04-17 18:11:51 UTC
Now visiting this bug again, I found that there is no issues in the logs, and server side behavior. This can happen because connection is not established when the reply was about to be sent. The ping-pong hanging issue should be related to ping-timeout not properly done or something else. Surely not an issue on the server side.

Re-open with 'ping-pong tests hang on fuse mount' summary if the behavior persist on latest mountpoint. Moving to 'wontfix' status for now.