Re: [wmii] wmiirc Locking Up

On Wed, Jun 14, 2006 at 07:02:38AM -0600, Ryan Kirkpatrick wrote:
> On Mon, 12 Jun 2006, Anselm R. Garbe wrote:
>
> > On Sat, Jun 10, 2006 at 12:04:07PM -0600, Ryan Kirkpatrick wrote:
> > > I have been using wmii for the past few weeks, and overall I have been
> > > quite happy with it. But I have stumbled across a rather annoying bug with
> > > wmiirc. After I have been running wmii for anywhere between a day and
> > > several days, it will stop responding to keyboard inputs. I blame this on
> > > wmiirc as if I kill it and then start a new one from an open shell, then
> > > keyboard input starts working again. This has happened multiple times.
> >
> > Never noticed this. Do you use AFS?
>
> No, only NFS, and not on the systems having this problem. My laptop has
> displayed it once or twice, and it has never has any network mounts. Where
> I have seen it more often is on my server, which is hosting several
> X-terminals. I have a couple of disk-less PCs running only an X server and
> using XDM to connect back to the server. All applications, including wmii,
> are running on the server, as X clients. While the server does offer NFS
> shares to other systems, the home directories that wmiirc is running out
> of are local on the server itself. Maybe it has something to do with
> multiple instances of wmiirc running on the same system at the same time?
>
> Here are a list of the running processes related to wmii, where 'magellan'
> is the server and furies, juno, vesta, hathaway, and apollo are the
> Xterminals. My user, rkirkpat, is logged on at each terminal. The
> /home/users/rkirkpat directory is local to magellan.
>
> magellan$ ps ax | grep wmii
> 903 ? S 0:00 /bin/sh /usr/bin/wmii
> 948 ? S 5:49 wmiiwm -a unix /tmp/ns.rkirkpat.furies.rkirkpat.net:0/wmii
> 972 ? S 20:04 /bin/sh /home/users/rkirkpat/.wmii-3/status
> 985 ? S 0:00 wmiir read /event
> 2921 ? S 0:00 /bin/sh /usr/bin/wmii
> 2966 ? S 5:51 wmiiwm -a unix /tmp/ns.rkirkpat.juno.rkirkpat.net:0/wmii
> 2990 ? S 20:14 /bin/sh /home/users/rkirkpat/.wmii-3/status
> 3005 ? S 0:00 wmiir read /event
> 14058 ? S 0:00 /bin/sh /usr/bin/wmii
> 14103 ? S 6:29 wmiiwm -a unix /tmp/ns.rkirkpat.vesta.rkirkpat.net:0/wmii
> 988 ? S 0:00 /bin/sh /home/users/rkirkpat/.wmii-3/wmiirc
> 3006 ? S 0:00 /bin/sh /home/users/rkirkpat/.wmii-3/wmiirc
> 32541 ? S 0:00 /bin/sh /usr/bin/wmii
> 32607 ? S 2:43 wmiiwm -a unix /tmp/ns.rkirkpat.hathaway.rkirkpat.net:0/wmii
> 32641 ? S 9:06 /bin/sh /home/users/rkirkpat/.wmii-3/status
> 32656 ? S 0:00 wmiir read /event
> 30541 ? S 5:20 /bin/sh /home/users/rkirkpat/.wmii-3/status
> 30553 ? S 0:00 wmiir read /event
> 30554 ? S 0:00 /bin/sh /home/users/rkirkpat/.wmii-3/wmiirc
> 10194 ? S 0:00 /bin/sh /usr/bin/wmii
> 10262 ? S 1:37 wmiiwm -a unix /tmp/ns.rkirkpat.apollo.rkirkpat.net:0/wmii
> 10316 ? S 5:28 /bin/sh /home/users/rkirkpat/.wmii-3/status
> 10335 ? S 0:00 wmiir read /event
> 10336 ? S 0:00 /bin/sh /home/users/rkirkpat/.wmii-3/wmiirc
> 32657 ? S 0:00 /bin/sh /home/users/rkirkpat/.wmii-3/wmiirc
>
> I am still running straces on the wmiirc processes to try and catch the
> lock up in the act. TTYL.

I doubt that running multiple wmiirc's connected to different
wmiiwm instances running on different servers might be a reason
for this behavior (I run several X servers as well).
The process list of magellan looks like /tmp is exported among
all diskless clients. This also means that the socket file might
be accessed through NFS, which might cause trouble.

Regards,

-- 
 Anselm R. Garbe  ><><  www.ebrag.de  ><><  GPG key: 0D73F361