[indiana-discuss] evince hangs in b101 after suspend-resume

[Jürgen Keil]

> solarg wrote:
> > Jürgen Keil wrote:
> >> Does a truss with option "-v all" reveal the IP address and TCP port number
> >> that we're trying to connect to?
> >>
> the problem arises now when i'm trying to go into submenus 
> of evince, and truss reveals the same port and unix socket concerned:
>
> /1: connect(23, 0x087DD648, 48, SOV_DEFAULT) Err#146 ECONNREFUSED
> /1: AF_UNIX name =/var/tmp/orbit-henry/linc-2d4-0-49a39032c7fd8

The failed unix socket connect reveals more information:

The hexadecimal number following ".../orbit-$USER/linc-"
("2d4" in this case) should be PID of process that has created
the socket.


Converting 2d4 to decimal is PID 724:

% echo 2d4=D | mdb
                724 


So it should be PID 724 that is supposed to listen and
accept connections with that unix domain socket address.


The next time the system has problems on resume, try to
"truss -v all" the hanging gnome application, try to find one
of those failed unix domain socket connects, convert the
hexadecimal number following the "linc-" substring to a 
decimal PID value, and run "pstack $PID".

Now it is possible that the process with $PID does not run any
more (has crashed) after such a failed suspend / resume.  It
would be best if you could save the output of "ptree" everytime
before you suspend the system, so that you can search which
process did use $PID before the failed suspend /resume.
-- 
This message posted from opensolaris.org