sbdrop kernel panic

[rezba]

I had a kernel panic on a machine today, it was the first one on this unit since 10.1. was installed, in oct.2001.

the kp message was :

panic(cpu 0): sbdrop

and a common message for the latest stack backtrace for cpu 0, with only few lines without any relevant informations.

Before that, because of strange behaviours, I log out/in two times, with, each times, a brief command line sreen after logging out, showing problems with classic librarys in core services.

I had problems with classic yesterday. A FileMaker 5 version for 9 crashed my classic environment three times.

Before all, I decided to change the location of my Classic system, in the Classic Preferences, log out/login.

Then, not but funny messages had been displayed, such as :
no more space available in my OSX partition (was copying a 4kb file !)
impossible to write (in eudora)
impossible to backup your iTunes Library

and even : impossible to write out new records (4D server message, which database is stored on a 15GO free partition!).

Of course, I do have free space...

Just before the kp, I ran MOXOptimize to see how much RAM was used.
I have 1,2 Go RAM with this dual G4 533, and the only apps opened at this moment were Eudora, OW, terminal, 4D server, suitcase server, acrobat 5. Only carbonized or cocoa apps. But MOX only show me 15 free mg of ram, and I was paging out a bit (4000 pageout).

But, most of all, MOX showed me that the classic stratup items were enabled, for a amount of 40Mg. But I haven't launched classic in this session !

Log out. KP happens.

I reboot, run fsck -y
- minor catalog problems,
- major multi-linked file problem
and finally major volume information problem :
amount of block was 1 instead of 1335xxxxxxx !


Everything seems to be OK now, and I will check for a X versoin of Filemaker.

But I don't have any idea of the real causes of that.

Anyone, please ?

 

[blb]

Well, a quick search on the kernel shows that sbdrop() is Drop data from (the front of) a sockbuf, so it's in networking code when you hit the panic. I'd say you've either stumbled onto a bug or you have hardware that's flaking out on you.

 

[rezba]

right. apparently, what has crashed is the network setup component of the 4D server. This component is not carbonized yet, and 4D use it sometimes when he's loosing ping. Conjunction of something like that and the sudden instability of classic could explain this. Thanks blb.