| Title: | DECWINDOWS 26-JAN-89 to 29-NOV-90 |
| Notice: | See 1639.0 for VMS V5.3 kit; 2043.0 for 5.4 IFT kit |
| Moderator: | STAR::VATNE |
| Created: | Mon Oct 30 1989 |
| Last Modified: | Mon Dec 31 1990 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 3726 |
| Total number of notes: | 19516 |
Multiple create /term /disp=NODE::0 commands causes RWAST state
for the DECW$TE_0 server, we had problems with a MUTEX state earlier.
The MUTEX state was cured by adding some PQL_* parameters
in MODPARAMS.DAT (see below).
We have also tried some DECW$* logicals, but nothing seems to help.
define /table=DECW$LOGICAL_NAMES DECW$SERVER_WSEXTENT 10000
define /table=DECW$LOGICAL_NAMES DECW$DECTERM_CTRL_WSEXTENT 10000
define /table=DECW$LOGICAL_NAMES DECW$DECTERM_CTRL_WSQUOTA 10000
We are running VMS 5.3-1.
In order to set up a test environment, we need approx. 15 subprocesses.
What's the problem?????
/bod
The text below describes our problem:
$ create /term /disp=NODE::0 !!! starting the first subprocess
$ show system
VAX/VMS V5.3-1 on node BAKOM 19-SEP-1990 10:33:19.34 Uptime 4 19:46:58
Pid Process Name State Pri I/O CPU Page flts Ph.Mem
20600582 OMNI_TEST_1 LEF 4 158 0 00:00:00.37 197 190 S
:
2060054F DECW$TE_0 LEF 6 147 0 00:00:02.17 1726 1700
$ show /proc /quota /id=2060054F
19-SEP-1990 10:33:51.68 User: OMNI_TEST Process ID: 2060054F
Node: BAKOM Process name: "DECW$TE_0"
Process Quotas:
Account name: OMNI
CPU limit: Infinite Direct I/O limit: 18
Buffered I/O byte count quota: 14768 Buffered I/O limit: 18
Timer queue entry quota: 8 Open file quota: 13
Paging file quota: 18083 Subprocess quota: 20
Default page fault cluster: 64 AST quota: 42
Enqueue quota: 30 Shared file limit: 0
Max detached processes: 0 Max active jobs: 0
$ create /term /disp=NODE::0 !!! starting another subprocess
$ create /term /disp=NODE::0 !!! starting another subprocess
$ create /term /disp=NODE::0 !!! starting another subprocess
$ create /term /disp=NODE::0 !!! starting another subprocess
$ create /term /disp=NODE::0 !!! starting another subprocess
$ create /term /disp=NODE::0 !!! starting another subprocess
$ show /proc /quota /id=2060054F
19-SEP-1990 10:35:09.73 User: OMNI_TEST Process ID: 2060054F
Node: BAKOM Process name: "DECW$TE_0"
Process Quotas:
Account name: OMNI
CPU limit: Infinite Direct I/O limit: 18
Buffered I/O byte count quota: 11648 Buffered I/O limit: 18
Timer queue entry quota: 8 Open file quota: 13
Paging file quota: 16051 Subprocess quota: 20
Default page fault cluster: 64 AST quota: 22
Enqueue quota: 30 Shared file limit: 0
Max detached processes: 0 Max active jobs: 0
$ create /term /disp=NODE::0 !!! starting another subprocess
$ show system
VAX/VMS V5.3-1 on node BAKOM 19-SEP-1990 10:36:15.91 Uptime 4 19:49:55
Pid Process Name State Pri I/O CPU Page flts Ph.Mem
20600582 OMNI_TEST_1 LEF 4 158 0 00:00:00.37 197 190 S
20600503 OMNI_TEST_2 LEF 4 158 0 00:00:00.40 197 190 S
20600584 OMNI_TEST_3 LEF 4 158 0 00:00:00.38 197 190 S
20600505 OMNI_TEST_4 LEF 4 158 0 00:00:00.32 197 190 S
20600586 OMNI_TEST_5 LEF 4 158 0 00:00:00.46 197 190 S
2060059F OMNI_TEST_6 LEF 4 158 0 00:00:00.37 197 190 S
206005A0 OMNI_TEST_7 LEF 4 158 0 00:00:00.35 197 190 S
20600522 OMNI_TEST_8 LEF 4 158 0 00:00:00.39 197 190 S
:
2060054F DECW$TE_0 RWAST 7 742 0 00:00:08.68 4926 3191
================================================================================
MODPARAMS.DAT:
VMS5=1
SCSNODE="BAKOM"
SCSSYSTEMID=52013
ALLOCLASS=0
DUMPFILE=0 ! 891130 BOD
DUMPSTYLE=1
DISK_QUORUM=""
EXPECTED_VOTES=1
VOTES=1
INTERCONNECT="NI"
LOCKDIRWT=1
MSCP_LOAD=1
MSCP_SERVE_ALL=1
NISCS_CONV_BOOT=0
NISCS_LOAD_PEA0=1
NISCS_PORT_SERV=0
PAGEFILE=0 ! 900904 BOD
QDSKINTERVAL=10
QDSKVOTES=1
RECNXINTERVAL=20
STARTUP="SYS$SYSTEM:STARTUP_SITE.COM"
SWAPFILE=0 ! 900904 BOD
TTY_DEFCHAR2=%X21002
VAXCLUSTER=2
WINDOW_SYSTEM=0
!
! Layered products related parameters:
!
min_GBLPAGES=30000 ! 900911 BOD
min_GBLPAGFIL=7000 ! 890807 JH & BOD
min_GBLSECTIONS=500 ! 890710 JH & BOD
min_SRPCOUNT=2200 ! 890710 JH & BOD
min_IRPCOUNT=2000 ! 890710 JH & BOD
min_LRPCOUNT=50 ! 890710 JH & BOD
min_NPAGEDYN=900000 ! 890710 JH & BOD
min_VIRTUALPAGECNT=35000 ! 900907 KjS (For VTX)
!
add_GBLPAGES=2000 ! 891219 KjS (For RTR V1.0)
add_GBLPAGFIL=1600 ! 891219 KjS (For RTR V1.0)
add_GBLSECTIONS=20 ! 891219 KjS (For RTR V1.0)
add_SYSMWCNT=16 ! 891221 JH (For RTR V1.0)
!
PQL_DASTLM=48 ! 900914 BOD (For $ CREATE/TERMINAL...)
PQL_DBYTLM=20000 ! 900914 BOD (For $ CREATE/TERMINAL...)
PQL_DPGFLQUOTA=20000 ! 900914 BOD (For $ CREATE/TERMINAL...)
PQL_DPRCLM=20 ! 900914 BOD (For $ CREATE/TERMINAL...)
================================================================================
SYSGEN> SHO/PQL
Parameters in use: Active
Parameter Name Current Default Minimum Maximum Unit Dynamic
-------------- ------- ------- ------- ------- ---- -------
PQL_DASTLM 48 24 -1 -1 Ast D
PQL_MASTLM 4 4 -1 -1 Ast D
PQL_DBIOLM 18 18 -1 -1 I/O D
PQL_MBIOLM 4 4 -1 -1 I/O D
PQL_DBYTLM 20000 8192 -1 -1 Bytes D
PQL_MBYTLM 3000 1024 -1 -1 Bytes D
PQL_DCPULM 0 0 -1 -1 10Ms D
PQL_MCPULM 0 0 -1 -1 10Ms D
PQL_DDIOLM 18 18 -1 -1 I/O D
PQL_MDIOLM 4 4 -1 -1 I/O D
PQL_DFILLM 16 16 -1 -1 Files D
PQL_MFILLM 2 2 -1 -1 Files D
PQL_DPGFLQUOTA 20000 8192 -1 -1 Pages D
PQL_MPGFLQUOTA 512 512 -1 -1 Pages D
PQL_DPRCLM 20 8 -1 -1 Processes D
PQL_MPRCLM 0 0 -1 -1 Processes D
PQL_DTQELM 8 8 -1 -1 Timers D
PQL_MTQELM 0 0 -1 -1 Timers D
PQL_DWSDEFAULT 239 100 -1 -1 Pages
PQL_MWSDEFAULT 239 60 -1 -1 Pages
PQL_DWSQUOTA 478 200 -1 -1 Pages D
PQL_MWSQUOTA 478 60 -1 -1 Pages D
PQL_DWSEXTENT 478 400 -1 -1 Pages D
PQL_MWSEXTENT 478 60 -1 -1 Pages D
PQL_DENQLM 30 30 -1 -1 Locks D
PQL_MENQLM 4 4 -1 -1 Locks D
PQL_DJTQUOTA 1024 1024 -1 -1 Bytes D
PQL_MJTQUOTA 0 0 -1 -1 Bytes D
================================================================================
SYSUAF entry for user OMNI_TEST:
Maxjobs: 0 Fillm: 40 Bytlm: 36000
Maxacctjobs: 0 Shrfillm: 0 Pbytlm: 0
Maxdetach: 0 BIOlm: 18 JTquota: 1024
Prclm: 25 DIOlm: 18 WSdef: 1024
Prio: 4 ASTlm: 50 WSquo: 2048
Queprio: 0 TQElm: 20 WSextent: 8192
CPU: (none) Enqlm: 200 Pgflquo: 25000
================================================================================
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 3358.1 | Patch for the RWAST hang problem | HANNAH::MESSENGER | Bob Messenger | Wed Sep 19 1990 11:38 | 15 |
Re: .0 Dan CSC32::Butterworth has a patch for this problem (CLD CXO05901). Apparently what happens is that multiple messages are sent to a mailbox that only has enough room for one message. The patch is to increase the size of the mailbox. By the way, instead of create/term/display=node::0 you might find it more reliable to do this: $ set display/create/node=node $ create/term $ create/term ... etc. -- Bob | |||||
| 3358.2 | RWAST not RWMBX | VINO::MCARLETON | Reality; what a concept! | Wed Sep 19 1990 16:15 | 7 |
Re: .1
No Bob. Dan's patch will fix a TE process stuck in RWMBX state not
RWAST state. I think this is a different problem. (I thought the same
thing as you at first until I saw that it was RWAST not MBX.)
MJC
| |||||
| 3358.3 | me, too | DEMON3::CLEVELAND | Notes - fun or satanic cult? | Thu Sep 20 1990 09:22 | 5 |
I've seen the same thing (TE in RWAST). I found that the process was
waiting for a MUTEX in system space somewhere. I'll try to reproduce
it today...
Tim
| |||||
| 3358.4 | HANNAH::MESSENGER | Bob Messenger | Fri Sep 21 1990 10:17 | 5 | |
Re: .2 Oops... Since this is a new problem, someone should QAR it. -- Bob | |||||
| 3358.5 | New or not new, that's the question... | LNKHUB::DAHL | Bengt-Olof Dahl @LNK, EIS PA LNK | Mon Sep 24 1990 06:52 | 35 |
Re: .4
I found QAR #262 in the DECWINDOWS-IFT database.
Is it a new problem?
(Our MUTEX problem is described in QAR #4074 and #4138)
/bod
QAR # Status Sev Category Component Date in Date out
----- ------ --- -------- -------------- ------------ -------------
00262 CL H FN DECWTERM 13-JAN-1988 5-MAR-1988
BL6.4 TERMINAL EMULATOR RWASTS W 3 TERMINALS
Reproducible at will: Y
CPU Memory System device
VAX lots DEC
BL6.4 terminal Emulator: DECW$DECTERM goes into RWAST state
when I create 3 terminal emulator windows.
Answer for QAR #00262:
------ --- ------- -------
This problem can be avoided by increasing the BYTLM quota for the
terminal controller process. One way is to set PQL_DBYTLM to something
large like 50000. This could be corrected in the Session Manager, also.
The actual cause of the RWAST is within the TP/PYDRIVER, I believe.
--tom
DECterm's BYTLM requirements have been drastically reduced for FT1. You
should no longer need such a high BYTLM.
--tom
| |||||
| 3358.6 | DECWIN::MESSENGER | Bob Messenger | Tue Sep 25 1990 13:08 | 7 | |
Re: .6 Interesting; I knew that DECterm could hang in a MUTEX state when BYTLM was too low, but I didn't know that this could also result in an RWAST hang. It sounds like increasing BYTLM might fix the problem. -- Bob | |||||