printer problems PSM-E-WRITEERR

[Klaes-Göran Carlsson]

Hi all

The short version:
%PSM-E-WRITEERR, error writing !AS
-SYSTEM-F-HANGUP, data set hang-up

The long "detailed" version:

I have a lot of problems with my printers. Some are connected via tcp, others using old LAT. Guess the problem is not the printers itself.

Some jobs causing the printer queue to be stalled. The printfile is OK, it is printed out when reprinting it.
All printing is done from a C++ application using:
status = sys$sndjbcw(0,SJC$_ENTER_FILE,0,&item_list,0,0,0);
(I have no idea what that line is doing... just put it here so you don't have to ask:)

The application has been working for 10 years in 30 different factories, just has to be some quotas or network or what? I have adjusted "mcr sysgen sh CHANNELCNT" since autogen asked for it.

Guess your answers will be like "network related", (that's my answer to customers when I don't have any idea)
but... please give me a hint where to dig.

Thanks in advance
/Klaes-Goran


EINKGCA_BTRACK> ucx sh ver

HP TCP/IP Services for OpenVMS Alpha Version V5.4
on a AlphaServer 1200 5/533 4MB running OpenVMS V7.3-2




EINKGCA_BTRACK> sh que OBSPRINT142/all/full

Terminal queue OBSPRINT142, idle, on BTRACK::LTA629, mounted form DEFAULT

/AUTOSTART_ON=(BTRACK::LTA629) /BASE_PRIORITY=4 /DEFAULT=(FEED,FORM=DEFAULT)
Lowercase /OWNER=[SYSTEM] /PROTECTION=(S:M,O:D,G:R,W:S) /RETAIN=ERROR
/SCHEDULE=(NOSIZE)

Entry Jobname Username Blocks Status
----- ------- -------- ------ ------
4514 102_19002-2_SVV10799A23004H0JA
BAR_SFT 3 Retained on error
%PSM-E-WRITEERR, error writing !AS
-SYSTEM-F-HANGUP, data set hang-up
Submitted 21-NOV-2005 07:42:18.96 /FORM=DEFAULT /PRIORITY=100
File: _BTRACK$DRA1:[BARTRACK.062.TMP.LABEL]102_19002-2_SVV10799A23004H0
/DELETE
Completed 21-NOV-2005 07:42:18.99 on queue OBSPRINT142




EINKGCA_BTRACK> sh que z*215*/all/full
Printer queue ZBSPRINT215, idle, on BTRACK::"PSB202.bs.sw.ericsson.se:9100",
mounted form DEFAULT
/BASE_PRIORITY=4 /DEFAULT=(FEED,FORM=DEFAULT) Lowercase /OWNER=[SYSTEM]
/PROCESSOR=UCX$TELNETSYM /PROTECTION=(S:M,O:D,G:R,W:S) /RETAIN=ERROR

Entry Jobname Username Blocks Status
----- ------- -------- ------ ------
1817 4_19002-4_SVV10799A23004HGLY
BAR_SFT 19 Retained on error
%PSM-E-WRITEERR, error writing !AS
-SYSTEM-F-NOIOCHAN, no I/O channel available
Submitted 24-NOV-2005 09:35:00.77 /FORM=DEFAULT /PRIORITY=100
File: _BTRACK$DRA1:[BARTRACK.062.TMP.LABEL]4_19002-4_SVV10799A23004HGLY
/DELETE
Completed 24-NOV-2005 09:35:01.36 on queue ZBSPRINT215

[Klaes-Göran Carlsson]

Also have a lot of errors like:

1051 4_19002-4_SVV10799A23004J5KG
BAR_SFT 19 Retained on error
%PSM-E-WRITEERR, error writing !AS
-SYSTEM-F-NOIOCHAN, no I/O channel available
Completed 30-NOV-2005 07:56:41.46 on queue ZBSPRINT215
1300 4_19002-4_SVV10799A23004J5TP
BAR_SFT 19 Retained on error
%PSM-E-WRITEERR, error writing !AS
-SYSTEM-F-NOIOCHAN, no I/O channel available
Completed 30-NOV-2005 08:58:01.79 on queue ZBSPRINT215

[Uwe Zessin]

Hm, "dataset hang up" is indeed a sign of network disruptions, but NOIOCHAN most likely not. Apart from raising CHANNELCNT (which is not a dynamic parameter, so you need a reboot, but I think you know that).

It is possible that the C++ RTLs eats up too many I/O channels - does it help if you raise the system parameter PQL_MFILLM?

[Ian Miller.]

Raising CHANNELCNT and parhaps openfile quota for the symboint process should help with NOIOCHAN.

Dataset hangup from a lat connection indicates a network timeout. LAT has short timeouts, some of which can not be changed. Check for network issues between the DECserver and your VMS system.

[labadie_1]

Check the quota of the symbiont process that deals with your printer, with Amds or Availability Manager (both available at http://h71000.www7.hp.com/openvms/products/availman/index.html ) or pquota (http://h71000.www7.hp.com/freeware/freeware40/pquota/)

[Klaes-Göran Carlsson]

I have adjusted the PQL_MFILLM from 100 to 1000, have to let the system run for some hours before I know if this was the solution. Have been running 1 hour now without problem:)

Uwe: Yes, I have rebooted after changing channelcnt.

Ian: "open file quota" do you mean FILLM?

Then I need to learn some more: How to find out which process/account to adjust?

sh sys/proc=*symb* gives me a list of 5 processes. Show proc/id/cont tells me that some are running DCPS$SMB.EXE others running LATSYM.EXE
All of them with system account. Is it just to mcr authorize mod system/fillm=3000 ? It is =2000 now.



/Klaes-Göran

[labadie_1]

to find the symbiont for lta629, do
$ ana/sys
set proc/id=id_of_1st_symbiont
sh proc/chann
if you have lta629, it is this one, note the pid

[Uwe Zessin]

> Ian: "open file quota" do you mean FILLM?

I am not Ian, but I'd say yes ;-)


As far as I can tell, a symbiont is not activated in a way that takes care about a username's SYSUAF quotas - thus my suggestion to look after the system parameter PQL_MFILLM which overrules any lower values from the process quota list.

[labadie_1]

If memory serves me, a symbiont takes his quotas from jobctl process, so Uwe method is good and simple

[Ian Miller.]

Yes FILLM. Symbiont processes are detached processes so the values in the UAF do not apply. Raising PQL_MFILLM is the way to go.

[Klaes-Göran Carlsson]

It didn't helped to adjust pql_mfillm from 100 to 1000, does any of you have any more suggestions?


SYSGEN> USE ACTIVE
SYSGEN> SET PQL_MFILLM 1000
SYSGEN> WRITE ACTIVE
SYSGEN> EXIT

I have got 3 more printjobs with errors as below. I have downloaded and made some tests with PQUOTA, but since I'm a novice it doesn't tells me anything. Have run it on all 5 TCPIP$TNSn processes. The FILCNT is max used 17%, all other 0.


Printer queue OBSPRINT290, idle, on BTRACK::"psbs36.bs.sw.ericsson.se:3001",
mounted form DEFAULT
/BASE_PRIORITY=4 /DEFAULT=(FEED,FORM=DEFAULT) Lowercase /OWNER=[SYSTEM]
/PROCESSOR=UCX$TELNETSYM /PROTECTION=(S:M,O:D,G:R,W:S) /RETAIN=ERROR

1838 104_19002-4_SVV10799A23004J6ET
BAR_SFT 3 Retained on error
%PSM-E-WRITEERR, error writing !AS
-SYSTEM-F-LINKDISCON, network partner disconnected logical link
Submitted 30-NOV-2005 11:41:14.57 /FORM=DEFAULT /PRIORITY=100
File: _BTRACK$DRA1:[BARTRACK.062.TMP.LABEL]104_19002-4_SVV10799A23004J6
/DELETE
Completed 30-NOV-2005 11:41:14.72 on queue OBSPRINT290

1953 104_19002-4_SVV10799A23004J6H5
BAR_SFT 3 Retained on error
%PSM-E-WRITEERR, error writing !AS
-SYSTEM-F-NOIOCHAN, no I/O channel available
Submitted 30-NOV-2005 12:23:21.63 /FORM=DEFAULT /PRIORITY=100
File: _BTRACK$DRA1:[BARTRACK.062.TMP.LABEL]104_19002-4_SVV10799A23004J6
/DELETE
Completed 30-NOV-2005 12:23:21.85 on queue OBSPRINT290

2117 104_19002-4_SVV10799A23004J6TQ
BAR_SFT 3 Retained on error
%PSM-E-WRITEERR, error writing !AS
-SYSTEM-F-LINKDISCON, network partner disconnected logical link
Submitted 30-NOV-2005 13:16:43.76 /FORM=DEFAULT /PRIORITY=100
File: _BTRACK$DRA1:[BARTRACK.062.TMP.LABEL]104_19002-4_SVV10799A23004J6






EINKGCA_BTRACK> sh sys/proc=*tns*
OpenVMS V7.3-2 on node BTRACK 30-NOV-2005 15:25:47.96 Uptime 12 23:22:56
Pid Process Name State Pri I/O CPU Page flts Pages
00000422 TCPIP$TNS2 HIB 5 765514 0 00:04:34.27 493 410
00000423 TCPIP$TNS1 HIB 5 515322 0 00:02:53.93 466 384
000004C4 TCPIP$TNS3 HIB 5 1801884 0 00:10:47.49 378 295
000004C8 TCPIP$TNS4 HIB 5 694999 0 00:04:10.37 47529 248
000004CA TCPIP$TNS5 HIB 5 100546 0 00:00:30.05 115188 36

[Jan van den Ende]

Klaes-Göran,

Ian mentioned CHANNELCNT in passing, but HAVE you already raised it?

What is the current value?

Are you running any databases?

Proost.

Have one on me.

jpe

[Uwe Zessin]

Isn't CHANNELCNT per individual process?

[Jan van den Ende]

Uwe,

CHANNELCNT is a SYSGEN parameter (and regrettably, NOT dynamic)

Proost.

Have one on me.

jpe

[Uwe Zessin]

Sigh, of course that is a system parameter!

My point was that it affects the size of the channel control blocks array _for each process_. It is not a system-wide limit.

[Klaes-Göran Carlsson]

Hi

Yes, I have changed the channelcnt:
Parameter Name Current Default Min. Max. Unit Dynamic
-------------- ------- ------- ------- ------- ---- -------
CHANNELCNT 2400 256 31 65535 Channels

And also rebooted (one week ago), no change.

Jan: Yes, I have
RDB 7.1-04 database. (Not any ODBC, no sqlserver)
Bea Message Queue (DMQ) with some 100 clients connected using CLS... also with some problems...network???
Some JRE/RMI clients
and 2-3 other vms-nodes running without problem

Today I got inabout 50 new printjobs with errors like:
%PSM-E-WRITEERR followed by

-SYSTEM-F-LINKDISCON
-SYSTEM-F-NOIOCHAN
-SYSTEM-F-CONNECFAIL

most of them like this one:
BAR_SFT 3 Retained on error
%PSM-E-WRITEERR, error writing !AS
-SYSTEM-F-NOIOCHAN, no I/O channel available


Also, HP people are installing new printers. Changing from Intermec to Zebra printers, i.e from LAT to TCPIP
They have created 72 new queues for these new tcp printers.

There have been some "old vms-guys" looking into the problem, but... they arn't old enought:) (And I have only spent 10 years with vms commands like copy and delete, never been close to stuff like sysgen and autogen)

Thanks a lot for all your help!

/Klaes-Goran

[Klaes-Göran Carlsson]

Found below in result from autogen. Will it help to raise lockidtbl from 6000 to 50000? What a huge difference, why? Is a value = 50000 "normal" for lockidtbl?
Yes, I'm going to change it and reboot since its not dynamic.

LOCKIDTBL parameter information:
Feedback information.
Old value was 6000, New value is 46960
Current number of locks: 67898
Peak number of locks: 89088

RESHASHTBL parameter information:
Feedback information.
Old value was 4096, New value is 131072
Current number of resources: 47279


/Klaes-Goran

[Volker Halle]

Klas-Göran,

you first need to troubleshoot your printing problem based on the print-symbiont, which is controlling the print-queue, on which the error is happening (LATSYM or UCX$TELNETSYM).

TELNET print symbiont processes are called TCPIP$TNSx in recent versions of TCPIP. They should have a logfile in SYS$SPECIFIC:[SYSEXE]TCPIP$TELNETSYM_.LOG.

Once you find such a print-symbiont process in your system, issue a SDA> SHOW PROC/CHAN/ID= to find out how many channels it has. If the ni. of channels is near CHANNELCNT, this would explain the NOIOCHAN error.

See http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=941760 for a procedure to report the Symbiont PIDs for all the device queues...

Volker.

[Klaes-Göran Carlsson]

Hi All

It is network related! Suddenly some printer didn't reply on tcptraceroute. HP-guys will take care of it, not my problem any more.

I have learned a lot these days, thanks a lot!

/Klaes-Göran

[Klaes-Göran Carlsson]

closing