kjs%tufts.csnet@csnet-relay.arpa (Kevin Sullivan) (04/03/86)
Our 780 running 4.2 has been crashing with "panic: mchk," bet we are not getting crash dumps, which makes it difficult to see what is happening. On other kinds of crashes we do not have this problem. Any hints? Is there a way to manually force a dump before rebooting? kevin sullivan kjs%tufts@csnet-relay
msitd32@ms3.UUCP (Russ Brandon) (04/16/86)
In article <2321@brl-smoke.ARPA>, kjs%tufts.csnet@csnet-relay.arpa (Kevin Sullivan) writes: > > Our 780 running 4.2 has been crashing with "panic: mchk," bet we are not > getting crash dumps, which makes it difficult to see what is happening. > On other kinds of crashes we do not have this problem. Any hints? Is there > a way to manually force a dump before rebooting? > First check your /etc/rc file to see if contains a command '/etc/savecore directory_name'. This line causes a core image to be saved in the named directory when rebooting after a system crash. Also check to make sure that this directory is large enough to handle the core image. You space requirements will be (the size of /vmunix) + (the size of memory). example: if /vmunix is size 286720 on an 8Mbyte system, the amount of space needed in the directory would be (286720 + 8000000). You may want to dedicate an entire filesystem for this purpose. For futher info see savecore(8). -- Russell Brandon ...!seismo!vrdxhq!ms3!RSB c/o ISN, Crop 1235 Jefferson Davis Hwy Arlington, Va. 22202
smb@ulysses.UUCP (Steven Bellovin) (04/21/86)
> > Our 780 running 4.2 has been crashing with "panic: mchk," bet we are not > getting crash dumps, which makes it difficult to see what is happening. > On other kinds of crashes we do not have this problem. Any hints? Is there > a way to manually force a dump before rebooting? I've noticed that we don't get dumps from panics at interrupt level.
joe@ccnysci.UUCP (04/23/86)
> > Our 780 running 4.2 has been crashing with "panic: mchk," bet we are not > getting crash dumps, which makes it difficult to see what is happening. > On other kinds of crashes we do not have this problem. Any hints? Is there > a way to manually force a dump before rebooting? I would have to guess that your system is configured with ra81 or ra80 disk drives. After having the same problem and assuming for a long time that it was hardware related I discovered that the system core dump routine for the uda50 disk controller is a dummy routine: It sets the global errno variable to ENODEV and returns. By the time I discovered this the release of 4.3 seemed immiment ( !!!! ) and so I decided to await the possibility of a fix there. If anyone has a uda50 driver for 4.2 which includes a functional dump routine, I would appreciate getting it, as would many other people I'm sure. ihnp4!likewise!ccnysci!joe (Joe Eccles)