Process hangs 4ever in disk(?) wait w/negative priority on Sun4/SunOS

[Brent Southard]

I have a program that, after running for a few hours, somehow gets hung
with a negative priority.  Once in this state it remains so until I
reboot.  According to ps, it is in a disk (or other short term) wait, and
the WCHAN is Sysbase (this is on a SPARCstation 1 running SunOS 4.03).

Here is the relevant "ps lx" output:

       F UID   PID  PPID CP PRI NI  SZ  RSS WCHAN    STAT TT  TIME COMMAND
    8001 118 12628     1  0  -3  05544    0 Sysbase  D    q0114:11 foo
    8001 118 28240     1  0  -3  05476    0 Sysbase  D    p0 35:25 foo
    8000 118 28321     1204 -15  05476    0 Sysbase  DW   p0 35:42 foo

Does anyone out there have any ideas on what Sysbase is, or what could
cause a process to get stuck in this type of situation?  Also, does anyone
have any ideas on how to go about debugging this type of problem?  The
program does not always hang, and when it does it does not always do so at
the same point (i.e., sometimes it gets further along in its processing,
but it may be the same code that always causes the problem -- I don't
know).  Running under a debugger is useless, since once the problem occurs
I can't stop the program because it is immune to signals due to its
negative priority.

Any help would be appreciated.  Please post or reply to:

    scott at thirdi
    {pyramid,apple,hpda,sun}!thirdi!scott

Scott Southard
Third Eye Software