Mac Geekery

I just ran into this and Googled and found nothing but questions. I’ll go over what happened to me and how I resolved it. Maybe this will help others who run into this.

I installed all of the recommended Security Updates today (one of which, QuickTime 7.1, solves those image buffer overflow bugs) and after the reboot I noticed that the login panel wouldn’t let me type. While kind of odd, a reboot solved that. Then I noticed some background CPU activity, but I wrote that off as Spotlight doing something annoying again. In hindsight, these were part of the problem, but at the time could be dismissed as silly little things.

The fun began when I tried to start a program later tonight and got an error that said that the system was unable to start the program because of a -10810 error. I found the program, fiddled with it a little (even trying Rosetta) and nothing worked. Then I tried another program and the same thing happened. Well, looking that up was kind of futile but put me on the right track. I went to Apple’s Developer site and searched on “error 10810” (don’t put in the negative, that removes the word from the search). The result was the Launch Services error code list which called it an unknown error. Well, at least it was Launch Services and not the file.

So I looked at the system.log file and lo, tons of errors.

May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full
May 12 00:19:24 Kalyke kernel[0]: proc: table is full

Skipping around, I also found:

May 12 00:20:20 Kalyke /usr/libexec/bootpd: bind call failed: Address already in use
May 12 00:20:20 Kalyke /usr/libexec/bootpd: bind call failed: Address already in use
May 12 00:20:20 Kalyke /usr/libexec/bootpd: bind call failed: Address already in use
May 12 00:20:20 Kalyke /usr/libexec/bootpd: bind call failed: Address already in use
May 12 00:20:20 Kalyke /usr/libexec/bootpd: bind call failed: Address already in use
May 12 00:20:20 Kalyke /usr/libexec/bootpd: bind call failed: Address already in use
May 12 00:20:20 Kalyke /usr/libexec/bootpd: bind call failed: Address already in use
May 12 00:20:20 Kalyke /usr/libexec/bootpd: bind call failed: Address already in use

So, for me, it looks like bootpd was getting hung up when starting and the system kept trying to start it for some reason, filling up the process table and making it so that nothing else could start. Going back in the log, I found the same thing happening when the login window was horked. A simple sudo service bootps stop stopped the failure and then programs would start again. In the end, the unknown error was nothing more than all of the process table entries being in use.