[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference smurf::dec_mls_plus

Title:dec_mls_plus
Moderator:SMURF::BAT
Created:Mon Nov 29 1993
Last Modified:Thu Jun 05 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:534
Total number of notes:2544

513.0. "ORACLE: network operations cannot be killed on V4" by SMURF::BAT (Segui la tua beatitudine) Fri May 16 1997 19:04

From:	US2RMC::"[email protected]" "CWOLF.US.ORACLE.COM" 16-MAY-1997 15:56:11.57
To:	smurf::bat, [email protected]
CC:	[email protected]
Subj:	network operations can't be killed on MLS+ 4.0

Hello Barbara, 
 
We now have a problem which we were not seeing on MLS+ 3.1.  Which is that 
if we issue a network command that hangs, such as telnet, ping or ftp,  
and then try to kill it from another window, or abort it (<cntrl> '\'), 
it doesn't respond.  The only way out of it is to kill the parent process.  
 
It's been my experience that this occurs when the kernel software calls 
the kernel sleep routine at too high of a priority.  Any chance of getting 
this looked at? 
 
Thanks, 
 
	-Chris 
 
+------------------------------------------------------------------------+ 
| ORACLE Corporation                                                     | 
|                                                  [email protected]   | 
| Chris Wolf, Senior Software Engineer             Voice +1.415.506.2529 | 
| World Wide Government & Education                Fax   +1.415.506.7800 | 
| Products Division                                                      | 
+------------------------------------------------------------------------+
T.RTitleUserPersonal
Name
DateLines
513.1why is it hanging... is it really hanging?SMURF::BATSegui la tua beatitudineFri May 16 1997 19:0535
From:	SMURF::BAT          "Barbara A. Thomson ZKO3-2/X46 1-2955" 16-MAY-1997 18:03:07.10
To:	US2RMC::"[email protected]"
CC:	BAT
Subj:	RE: network operations can't be killed on MLS+ 4.0


	Yes, of course you can have this looked at :-)

	Can you give me a clue why the telnet/ping/ftp is hanging?
	E.g., is it because you haven't yet configured the object,
	or the object (host or interface) is configured incorrectly,
	or is it just a matter of it's the first time you're talking
	to the given host and it is doing its "tag mapping" ritual
	which takes an age, and it is timing out or something, or
	you want to kill it before it times itself out? (2 minutes
	or something I cannot remember what the timeout interval is.)

	(If you turn on tnet debugging with tnetd_ctl to -D2 level
	at least for at least the map daemon or the rh daemon,
	e.g., /tcb/bin/tnetd_ctl -D2 tnmapd ; /tcb/bin/tnetd_ctl -D2 tnrhd )
	and tail the log files in /var/adm/tnet/logs (by default),
	is anything interesting appearing there?

	I have a vague memory of once looking at the vanilla DU 
	login code and finding that there was a "hang" potential
	that was actually in the shell<->login interaction -- if
	the rlogin/whatever died, the shell would appear to be
	hanging waiting for something not there to return... but
	I'm assuming that this is not that... right?

	P.S. Whatever happened re: the ifdef that is not supposed
	to be there?

	Thanks
	BAT
513.2yes it hangsSMURF::BATSegui la tua beatitudineMon May 19 1997 20:1515
From:	US2RMC::"[email protected]" "CWOLF.US.ORACLE.COM" 16-MAY-1997 21:08:13.56
To:	smurf::bat
CC:	[email protected]
Subj:	Re: RE: network operations can't be killed on MLS+ 4.0

	[ excerpted ]
 
>you want to kill it before it times itself out? (2 minutes 
>or something I cannot remember what the timeout interval is.) 
 
Right, it's not important why a particular operation hangs -- it's just  
that we want to be able to kill it.  e.g. We forget to setup a host and  
then try to ping it.  Ping hangs.  On 3.2 we can just do <ctrl>'C' (SIGINT) 
to get out of it, but not on 4.0. 
 
513.3tnrhd is hanging: cannot -e /dev/spd7SMURF::BATSegui la tua beatitudineMon May 19 1997 22:0216
    Chris called and said he had things like
    
    	automount
    	tnet_kstats
    	ping
    
    in hangs that he cannot kill.  I had him dbx -k /vmunix set
    set $pid=those_pids and do a record output filename and t
    on each one.  He'll send me the trace.
    
    It appears that sec_find_rh is the problem, and that it is the
    tnrhd daemon that is hung.  I had him kill it -- it wouldn't
    kill with tnet_ctl -d tnrhd.  There is nothing in
    /var/adm/tnet/logs/tnrhd.log.  I asked if TNETRHDB is messed up.
    
    He's going to delete TNETDB and reboot.
513.4sec_find_rhSMURF::BATSegui la tua beatitudineTue May 20 1997 13:28109
Date: 19 May 97 21:03:51 -0400
From: "CWOLF.US.ORACLE.COM" <[email protected]>
To: [email protected], [email protected]
Subject: Hung network processes

Barbara, 
 
Just to let you; I deleted TNETDB and rebooted.  Now everything works, 
including the wildcarding of unlabled hosts.  In any case here are the  
states of those hung processes:  
 
===== trace of ping ================ 
 
[2] record output hang.trc (0 lines) 
(dbx)  
(dbx)  
>  0 thread_block() ["../../../../src/kernel/kern/sched_prim.c":2063, 
0xfffffc00002ac7ac] 
   1 thread_sleep(event = 0, lock = 0xfffffc0002126cd8, interruptible = 0) 
["../../../../src/kernel/kern/sched_prim.c":1735, 0xfffffc00002ac0b4] 
   2 spd_send_message(0xffffffff880af600, 0x120, 0x0, 0x0, 0xfffffc0007030120) 
["../../../../src/kernel/sec/sec_driver.c":1088, 0xfffffc00004f756c] 
   3 sec_find_rh(0x0, 0xfffffc0001f6e500, 0x0, 0x9, 0xfffffc0000000030) 
["../../../../src/kernel/sec/sec_tnet.c":1459, 0xfffffc00004fa85c] 
   4 tnsmm_proto_by_addr(0x0, 0x9, 0xfffffc0000000030, 0xfffffc00073c3400, 
0xfffffc00005093d0) ["../../../../src/kernel/tnet/tnsmm_protosw.c":113, 
0xfffffc00005073e8] 
   5 tnsmm_session_initiate(0xfffffc0005dd0f00, 0xfffffc0001f6e500, 0x0, 
0xfffffc00073c34b8, 0x0) ["../../../../src/kernel/tnet/tnsmm_socket.c":1991, 
0xfffffc00005093cc] 
   6 tnsmm_export_attr(0xfffffc0001bd2400, 0x0, 0xfffffc0001f6f700, 
0xfffffc00073c34b8, 0xfffffc0001f6f700) 
["../../../../src/kernel/tnet/tnsmm_socket.c":1680, 0xfffffc0000508e20] 
   7 tnsmm_socket_send(0x1fc0, 0x1, 0x40, 0x0, 0xfffffc0001f6e500) 
["../../../../src/kernel/tnet/tnsmm_socket.c":1087, 0xfffffc00005082bc] 
   8 sosend(0x0, 0x100000040, 0x0, 0xfffffc00073c34b8, 0xfffffc0000000000) 
["../../../../src/kernel/bsd/uipc_socket.c":1339, 0xfffffc0000292a44] 
   9 sendit(0xfffffc0001f6e500, 0xffffffff880af7f8, 0x40, 0x1, 0x100000000) 
["../../../../src/kernel/bsd/uipc_syscalls.c":910, 0xfffffc0000297158] 
  10 sendto(0x10, 0xffffffff880af7f8, 0xfffffc0000000001, 0x0, 0x2) 
["../../../../src/kernel/bsd/uipc_syscalls.c":749, 0xfffffc0000296d54] 
More (n if no)? 
========= trace of showmount =============== 
(dbx) warning: thread 0xfffffc00013ba2c0 appears to be swapped out 
(dbx) warning: thread 0xfffffc00013ba2c0 appears to be swapped out 
warning: thread 0xfffffc00013ba2c0 appears to be swapped out 
warning: thread 0xfffffc00013ba2c0 appears to be swapped out 
warning: thread 0xfffffc00013ba2c0 appears to be swapped out 
>  0 thread_block() ["../../../../src/kernel/kern/sched_prim.c":2063, 
0xfffffc00002ac7ac] 
warning: thread 0xfffffc00013ba2c0 appears to be swapped out 
warning: thread 0xfffffc00013ba2c0 appears to be swapped out 
   1 thread_sleep(event = warning: thread 0xfffffc00013ba2c0 appears to be 
swapped out 
0, lock = 0xfffffc0002126cd8, interruptible = warning: thread 
0xfffffc00013ba2c0 appears to be swapped out 
0) ["../../../../src/kernel/kern/sched_prim.c":1735, 0xfffffc00002ac0b4] 
   2 spd_send_message(0xffffffff881db600, 0x120, 0x0, 0x0, 0xfffffc0007030120) 
["../../../../src/kernel/sec/sec_driver.c":1088, 0xfffffc00004f756c] 
   3 sec_find_rh(0x0, 0xfffffc0003f78300, 0x0, 0x9, 0xfffffc0000000030) 
["../../../../src/kernel/sec/sec_tnet.c":1459, 0xfffffc00004fa85c] 
   4 tnsmm_proto_by_addr(0x0, 0x9, 0xfffffc0000000030, 0xfffffc00027d4000, 
0xfffffc00005093d0) ["../../../../src/kernel/tnet/tnsmm_protosw.c":113, 
0xfffffc00005073e8] 
   5 tnsmm_session_initiate(0xfffffc0005dd0180, 0xfffffc0003f78300, 0x0, 
0xfffffc00027d40b8, 0x0) ["../../../../src/kernel/tnet/tnsmm_socket.c":1991, 
0xfffffc00005093cc] 
   6 tnsmm_export_attr(0xfffffc0003f78a00, 0x0, 0xfffffc0003f79a00, 
0xfffffc00027d40b8, 0xfffffc0003f79a00) 
["../../../../src/kernel/tnet/tnsmm_socket.c":1680, 
More (n if no)?(dbx) warning: thread 0xfffffc0005b15340 appears to be swapped 
out 
(dbx)  
 
========= trace of tnetd_ctl =============== 
 
warning: thread 0xfffffc0005b15340 appears to be swapped out 
warning: thread 0xfffffc0005b15340 appears to be swapped out 
warning: thread 0xfffffc0005b15340 appears to be swapped out 
warning: thread 0xfffffc0005b15340 appears to be swapped out 
>0 thread_block() ["../../../../src/kernel/kern/sched_prim.c":2063, 
0xfffffc00002ac7ac] 
warning: thread 0xfffffc0005b15340 appears to be swapped out 
warning: thread 0xfffffc0005b15340 appears to be swapped out 
1 thread_sleep(event = warning: thread 0xfffffc0005b15340 appears to be 
swapped out 
0, lock = 0xfffffc0006fe9e18, interruptible = warning: thread 
0xfffffc0005b15340 appears to be swapped out 
0) ["../../../../src/kernel/kern/sched_prim.c":1735, 0xfffffc00002ac0b4] 
2 spdread(0x0, 0x0, 0xfffffc000062efb0, 0xfffffc0006fe9e00, 0x501200027) 
["../../../../src/kernel/sec/sec_driver.c":1454, 0xfffffc00004f7d88] 
3 spec_read(0xfffffc000049c6ec, 0xfffffc0003123800, 0xfffffc000049c688, 
0xfffffc0003123800, 0xffffffff01200027) 
["../../../../src/kernel/vfs/spec_vnops.c":1975, 0xfffffc000049c950] 
4 spec_clread(0xfffffc00004beccc, 0x0, 0xffffffff883cb818, 0xfffffc0003123800, 
0xfffffc000645d8c0) ["../../../../src/kernel/vfs/spec_vnops.c":1864, 
0xfffffc000049c698] 
5 vn_read(0xfffffc000028d534, 0xffffffff883cb8b0, 0xffffffff883cb818, 0x0, 
0xfffffc0001653100) ["../../../../src/kernel/vfs/vfs_vnops.c":926, 
0xfffffc00004becc8]More (n if no)?More (n if no)?    
6 rwuio(0xfffffc000133b720, 0xfffffc0005b15340, 0xfffffc00006a99a0, 
0xfffffc000133b500, 0xfffffc0001653100) 
["../../../../src/kernel/bsd/sys_generic.c":1217, 0xfffffc000028d588] 
   7 read(0x100010388, 0xffffffff883cb808, 0x40, 0x1038800000001, 0x0) 
["../../../../src/kernel/bsd/sys_generic.c":1167, 0xfffffc000028d498] 
   8 syscall(0x10388, 0x11ffffca8, 0x5, 0x11ffffca0, 0x3) 
["../../../../src/kernel/arch/alpha/syscall_trap.c":547, 0xfffffc0000523b80] 
   9 _Xsyscall(0x8, 0x3ff800d18e8, 0x1400098a0, 0x3, 0x140001fd0) 
["../../../../src/kernel/arch/alpha/locore.s":1177, 0xfffffc000051aca4]