I have two odd bugs happening. both happened about the time of a yum install so I assume they are related.


I was trying to instal rapidSVN that only has intranet (not internet) access. My defualt repository didn't have the rpm so I added what looked to be a good repository and installed rapidsvn. I then killed my vncserver connection with the intent to restart it so gnome would reflect any GUI changes caused by rapidSVN.

Unfortunately, my X11 client dies when I try to start vncserver now. There are many error messages, but most of the messages are the same as everyone got when they first connected to the server. The following messages are the *new* ones I get that aren't present in old functionaly VNC startups; written by hand so forgive typos:

______________-
Gtk-Message: Failed to load module "canberra-gtk-module": libcanberra-gtk-module.so: cannot open shared object file: no such file or directory
Gtk-Message: Failed to load module "canberra-gtk-module": libcanberra-gtk-module.so: cannot open shared object file: no such file or directory
Gtk-Message: Failed to load module "canberra-gtk-module": libcanberra-gtk-module.so: cannot open shared object file: no such file or directory
Gtk-Message: Failed to load module "canberra-gtk-module": libcanberra-gtk-module.so: cannot open shared object file: no such file or directory
Traceback (most recent call last):
File "/user/share/system-config-printer/applet.py", line 20, in <module>
inport cups
ImportError: /user/local/lib/libgcrypt.so.11: undefined symbol: gpg_err_set_errno
Failed to play sound: IO error

.
.
.
** (gnome-panel:24558): CRITICAL **: panel_applet_frame_change_background: assertion 'PANEL_IS_WIDGE (GTK_WDIGET (frame)->parent)' failed
/user/bin/Xvnc symol lokup error: /user/local/lib/libgcrypt.so.11: undefined symbole: gpg_err-set_errno
X10: fatal IO error 11( resource temporarily unavailable on X server ":6.0"
after 168 requests (168 known processed) with 0 events remaining.
gnome-pane: fatal IO error 104 (connection reset by peer) on X server :6.0.
gnome-volume-control-applet: Fatal IO error 11 (resource temporarily unavailable) on x server :6.0
... (4 more lines fo same)
The application 'nautilus' lost its connection to the siaply :6.0;
most likely the X server was shut down or killed/destoryed the application.

_______________-

I've also discovered that I can't do ps -ef, if I do the program will print the expected ps results, then hang indefinately. ps works fine though.


I'm not sure how to address this issue. I yum removed all four RPM installed when I tried to do rapidsvn, that didn't fix it. I don't know for sure that the yum install was even the problem; though it seems probable. Should I be looking at the gpg_err_set_errno undefined symbol or the missing camberra_gtk-module as the root cause of my problem? there is no camberra_gtk_module anywhere on my system according to locate...

Any help?

update/bump here

The ps command hanging issue went away with a reboot. quick search suggests that some pid out there had an issue where ps couldn't read details about it, so ps would block while trying to get specifics to output.

The other isues still exist. In fact the gpg_err_set_errno issue appears elsewhere now. I've found that if I just run gpg_auth itself I get the same error message. None of the gpg_auth files even appear to exist on my system. I've tried doing a yum install of gpg-auth-error (name may be slightly wrong, but error library for gpg-auth); but that didn't fix anything.

For now I want to focus on gpg-auth issue, though any information on the missing camberra-gtk-module would still be welcome. Any idea how I solve why gpg-auth is so broken for me ATM.

update.

I solved this. the issue was simple broken path. I had to set my LD_LIBRARY_PATH variable to point to /lib64; which is where libgpg_error .so was stored. gpg-agent new how to find the file fine, according to ldd. However, the libgcrypt.so; which is also a dependency of gpg-agent and run by it, was unaware of the location. I created a .sh file in /etc/profil.d which sets LD_LIBRARY_PATH to $LD_LIBRARY_PATH;/lib64 which should fix the problem perminatenly? still waiting on a reboot which I can't run now; but the real issue is resolved.

Incidentally; I would love to know how things got broken. my LD_LIBRARY_PATH is completely empty at bootup right now. shouldn't there be other locations LD_LIBRARY_PATH points to? am I going to run into something else breaking later because it's using LD_LIBRARY_PATH to find a library that isn't located in /lib64? should I manually add some other common directories to this path? or maybe figure out why it's empty to begin with?

Also, did my YUM install break this? I know yum didn't update libgpg_error. could yum have wiped out my LD_LIBRARY_PATH? done something else? I would love to believe that I was not the one that broke this; any chance YUM wasn't the fault lol.