when pantasks server shuts down pcontrol and pclient often dump core

[bills]

when pantasks server shuts down pcontrol and pclient often dump core.

Sample pcontrol stack trace

(gdb) where
#0 0x00002b51943143c5 in raise () from /lib64/libc.so.6
#1 0x00002b519431573e in abort () from /lib64/libc.so.6
#2 0x00002b51928a9205 in gprint (dest=GP_ERR, format=0x40b023 "\n")

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/lib.shell/gprint.c:300

#3 0x00000000004088ac in DownHosts ()

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/pcontrol/StopHosts.c:62

#4 0x0000000000403bb5 in cleanup ()

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/pcontrol/pcontrol.c:72

#5 0x00002b5192690b81 in quit (argc=1, argv=0x2aaab5fa76c0)

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/cmd.basic/quit.c:7

#6 0x00002b51928a3b7c in command (line=0x2aaab5fa7750 "quit",

outline=0x7fff18cd4840, VERBOSE=1)
at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/lib.shell/command.c:46

#7 0x00002b51928ab5f2 in multicommand (line=0x2aaab3941930 "quit")

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/lib.shell/multicommand.c:79

#8 0x00002b51928b54d0 in opihi (argc=1, argv=0x7fff18cd49d8)

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/lib.shell/opihi.c:42

#9 0x0000000000403c11 in main (argc=1, argv=0x7fff18cd49d8)

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/pcontrol/pcontrol.c:90

the assert is in gprint

297 if (stream[0].mode == GP_FILE) {
298 status = vfprintf (stream[0].file, format, argp);
299 if (status < 0) {
300 abort();
301 }
302 } else {
303 vPrintIOBuffer (stream[0].buffer, format, argp);
304 }

Sample pclient stack trace

(gdb) where
#0 0x00002b9c00ec53c5 in raise () from /lib64/libc.so.6
#1 0x00002b9c00ec673e in abort () from /lib64/libc.so.6
#2 0x0000000000401c66 in CheckChild ()

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/pclient/ChildOps.c:53

#3 0x00002b9bfff0359f in rl_read_key () from /lib64/libreadline.so.5
#4 0x00002b9bffef1d21 in readline_internal_char ()

from /lib64/libreadline.so.5

#5 0x00002b9bffef21a5 in readline () from /lib64/libreadline.so.5
#6 0x00002b9bfeb3a42e in opihi (argc=1, argv=0x7fffac3b3c08)

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/lib.shell/opihi.c:21

#7 0x0000000000401b05 in main (argc=1, argv=0x7fffac3b3c08)

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/pclient/pclient.c:77

(gdb) up
#1 0x00002b9c00ec673e in abort () from /lib64/libc.so.6
(gdb) up
#2 0x0000000000401c66 in CheckChild ()

at /data/ipp004.0/home/bills/ipp/Ohana/src/opihi/pclient/ChildOps.c:53

53 abort ();
(gdb) list 50
45 case, the opihi shell catches if the ssh dies using getppid
46 */
47 gettimeofday (&now, (void *) NULL);
48 dtime = 1e6*DTIME (now, last);
49 if (dtime < 50) {
50 Nbad ++;
51 if (Nbad > 100) {
52 gprint (GP_ERR, "serious IO error\n");
53 abort ();
54 }
(gdb) p Nbad
$1 = 101

[eugene]

this was caused by an attempt to write to a now-absent resource. I think this is fixed.