unkillable processes

From: Marty Leisner (mleisner@eng.mc.xerox.com)
Date: Fri Jun 07 2002 - 12:36:34 EDT


I'm running solaris 2.8.

I have old, unused processes I can't seem to kill as root.
bash2 :2 root@santa 11:46:19; ps -flL -p 12001
 F S UID PID PPID LWP NLWP C PRI NI ADDR SZ WCHAN STIME TTY LTIME CMD
 8 R amalik 12001 11963 1 1 0 40 20 ? 754 Apr 10 ? 0:00 writeint yy.int
bash2 :2 root@santa 11:46:44; kill -9 12001
bash2 :2 root@santa 11:46:54; ps -flL -p 12001
 F S UID PID PPID LWP NLWP C PRI NI ADDR SZ WCHAN STIME TTY LTIME CMD
 8 R amalik 12001 11963 1 1 0 40 20 ? 754 Apr 10 ? 0:00 writeint yy.int

bash2 :2 root@santa 11:48:34; pstack 12001
12001: writeint yy.int
 fee9ab08 read (0, ffbed8cf, 1)
 fefd4820 __getc (0, 0, 0, 0, 0, 0) + 20

bash2 :2 root@santa 11:50:16; pfiles 12001
pfiles: cannot control process 12001
bash2 :2 root@santa 11:50:33; lsof -p 12001
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
writeint 12001 amalik cwd VDIR 264,8192 1024 3014417 /apps/iqimages3/amal/SAGE/SIMULATION_amal/bin
writeint 12001 amalik txt VREG 264,388 25912 1561378 /apps/SP (earth-bay1:/e11/apps/SP)
writeint 12001 amalik txt VREG 85,10 24968 101137 /usr/lib/libmp.so.2
writeint 12001 amalik txt VREG 85,10 157032 101150 /usr/lib/libelf.so.1
writeint 12001 amalik txt VREG 85,10 884100 101139 /usr/lib/libnsl.so.1
writeint 12001 amalik txt VREG 85,10 382600 102398 /usr/lib/libresolv.so.2
writeint 12001 amalik txt VREG 85,10 1136744 101103 /usr/lib/libc.so.1
writeint 12001 amalik txt VREG 85,10 70792 102479 /usr/lib/libsocket.so.1
writeint 12001 amalik txt VREG 264,388 1671056 1398707 /apps/SP (earth-bay1:/e11/apps/SP)
writeint 12001 amalik txt VREG 85,10 78036 224295 /usr/ucblib/libucb.so.1
writeint 12001 amalik txt VREG 264,388 3321736 1398690 /apps/SP (earth-bay1:/e11/apps/SP)
writeint 12001 amalik txt VREG 264,388 77996 1398695 /apps/SP (earth-bay1:/e11/apps/SP)
writeint 12001 amalik txt VREG 85,10 102424 101202 /usr/lib/libm.so.1
writeint 12001 amalik txt VREG 85,10 4624 101146 /usr/lib/libdl.so.1
writeint 12001 amalik txt VREG 85,23 4500 11031 /var/ld/ld.config
writeint 12001 amalik txt VREG 85,10 196796 100903 /usr/lib/ld.so.1
writeint 12001 amalik 0u FIFO 0x3001458b020 0t0 1251185 (fifofs) PIPE->0x3001458b108
writeint 12001 amalik 1u VCHR 24,61 0t8458 114630 /devices/pseudo/pts@0:61->ttcompat->ldterm->ptem->pts
writeint 12001 amalik 2u VCHR 24,61 0t8458 114630 /devices/pseudo/pts@0:61->ttcompat->ldterm->ptem->pts
writeint 12001 amalik 3w VREG 264,8192 31224062 3014434 /apps/iqimages3 (dariat:/u05/apps/iqimages3)

pstop won't put the process into the T state...

I understand processes hung in a device wait, but there's no hardware here beyond files!!

This is a good question, I'll summarize responses (I've seen an answer from several years ago
to rm the /proc/<pid> files -- but I haven't tried that and I wonder about that!!).

I have dozens of processes like this (one is consuming 3gig of swap!!).

I can't believe I have to boot a machine to clear it!!

I'll summarize (and hopefully get some good answers which don't involve booting).

marty mleisner@eng.mc.xerox.com
Don't confuse education with schooling.
        Milton Friedman to Yogi Berra
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers



This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:24:25 EDT