DECWindows "White Screen" after VMS731_Graphics v4

[Rick Dyson]

I updated with some ECOs and now DECWindows does not startup correctly. It only gets to a "white" screen with the X11 mouse. No other response from mouse or keyboard except movement.

DECWindows was working fine before the updates. I have updated another node and it did not have this problem.

NOTE: After this initial work, I have since installed VMS731_UPDATE v6 *AND* VMS731_GRAPHICS v4, again after being suggested by HP Support.

Still just a white screen. I do not have Multinet nor have I ever done anything with Open3D. Both topics have been asked me by support. They don't have any answers yet...

Thanks!
Rick

Here are the ECOs I just installed initially:
sys_WOLF => product show history /sin=1-dec-2004
----------------------------------- ----------- -----------
--------------------
PRODUCT KIT TYPE OPERATION DATE AND
TIME
----------------------------------- ----------- -----------
--------------------
DEC AXPVMS VMS731_ACRTL V4.0 Patch Install 07-MAY-2005
14:53:40
DEC AXPVMS VMS731_GRAPHICS V4.0 Patch Install 07-MAY-2005
14:43:26
DEC AXPVMS VMS731_AUDSRV V2.0 Patch Install 07-MAY-2005
14:41:23
DEC AXPVMS VMS731_VMSMUP V1.0 Patch Install 07-MAY-2005
14:40:35
DEC AXPVMS DNVOSIMUP01 V7.3-1 Patch Install 07-MAY-2005
14:39:31
DEC AXPVMS DNVOSIECO03 V7.3-1 Patch Install 07-MAY-2005
14:37:57
DEC AXPVMS VMS731_RMS V6.0 Patch Install 12-FEB-2005
11:58:24
DEC AXPVMS VMS731_SYSINI V1.0 Patch Install 12-FEB-2005
11:57:25
DEC AXPVMS VMS731_UPDATE V5.0 Patch Install 12-FEB-2005
11:53:23
DEC AXPVMS VMS731_TRACE V3.0 Patch Install 12-DEC-2004
19:28:18
DEC AXPVMS VMS731_VERIFY V1.0 Patch Install 12-DEC-2004
19:27:37
DEC AXPVMS VMS731_LAN V12.0 Patch Install 12-DEC-2004
19:27:03
DEC AXPVMS VMS731_MAILSHR V1.0 Patch Install 12-DEC-2004
19:26:09
DEC AXPVMS VMS731_SHADOWING V3.0 Patch Install 12-DEC-2004
19:25:28
DEC AXPVMS VMS731_BACKUP V3.0 Patch Install 12-DEC-2004
19:24:23
DEC AXPVMS VMS731_IPC V2.0 Patch Install 12-DEC-2004
19:23:44
DEC AXPVMS VMS731_AUDSRV V1.0 Patch Install 12-DEC-2004
19:23:00
DEC AXPVMS VMS731_MANAGE V3.0 Patch Install 12-DEC-2004
19:22:18
DEC AXPVMS VMS731_F11X V4.0 Patch Install 12-DEC-2004
19:21:29
DEC AXPVMS VMS731_MOUNT96 V1.0 Patch Install 12-DEC-2004
19:20:27
DEC AXPVMS TCPIP_ECO V5.4-154 Patch Install 12-DEC-2004
09:31:47
----------------------------------- ----------- -----------
--------------------

21 items found



Here are the contents of Sys$Manager:DECW$Server_0_Error.log:

8-MAY-2005 11:14:28.1 Hello, this is the X server
This is the DECwindows X11 display server for OpenVMS AXP V7.3-050324
compiled on Mar 24 2005 at 08:41:34
Main address = 000386A0
Activating extension image DECW$SVEXT_Xie,
extension name: Xie, entry address 002E6780
Activating extension image DECW$SVEXT_DEC_XTRAP,
extension name: DEC-XTRAP, entry address 003B8548
%LIB-E-ACTIMAGE, error activating image DECW$SVEXT_MULTI_BUFFERING
-DSRINDEX-F-NOMSG, Message number 80C9FD84
DECW$XPORT_SERVICES image base address: 7C8B2000
DECW$TRANSPORT_DECNET image base address: 0045C000
%DECW-I-ATTACHED, transport DECNET attached to its network
DECW$TRANSPORT_LOCAL image base address: 0059E000
DECW$TRANSPORT_TCPIP image base address: 005E0000
%DECW-I-ATTACHED, transport TCPIP attached to its network
DECWINDOWS DigitalEquipmentCorp. AXP, Release 7.3
%LIB-E-ACTIMAGE, error activating image DECW$SERVER_DDX_GY

Unrecoverable server internal error (error code = 1381050) found, terminating al
l connections.
Mapped Images...

START END LENGTH IMAGE NAME
----- --- ------ ----------
10000 301ff 201ff DECW$SERVER_MAIN
32000 1e3bff 1b1bff DECW$SERVER_DIX
7bbac000 7bbddfff 31fff DECW$SECURITY_VMS
7b87a000 7b8ebfff 71fff SECURESHR
7b362000 7b3e3fff 81fff SECURESHRP
7c8c0000 7c8c9fff 9fff DECW$TRANSPORT_COMMON
7c8b8000 7c8bffff 7fff DECW$LCNLIBSHR
7c8ac000 7c8b7fff bfff DECW$XPORT_SERVICES
81420120 81432bf0 12ad0 SYS$BASE_IMAGE
7bfb6000 7c049fff 93fff DECC$SHR
7bc68000 7bcadfff 45fff DPML$SHR
7b91c000 7b92bfff ffff CMA$TIS_SHR
7b5b0000 7b601fff 51fff LIBRTL
7b602000 7b609fff 7fff LIBOTS
81403f58 81405f88 2030 SYS$PUBLIC_VECTORS
2e6000 3b73ff d13ff DECW$SVEXT_XIE
3b8000 3f83ff 403ff DECW$SVEXT_DEC_XTRAP
3fa000 43a3ff 403ff DECW$SVEXT_MULTI_BUFFERING
43c000 43ffff 3fff DECC$MSG
440000 4469ff 69ff SHRIMGMSG
448000 4581ff 101ff DECW$TRANSPORTMSG
45c000 59d5ff 1415ff DECW$TRANSPORT_DECNET
59e000 5de3ff 403ff DECW$TRANSPORT_LOCAL
5e0000 6203ff 403ff DECW$TRANSPORT_TCPIP
622000 7035ff e15ff DECW$SERVER_DDX_WMB_GY
82a000 88a9ff 609ff DECW$SERVER_DDX_CFB16
704000 7549ff 509ff DECW$SERVER_DDX_CFB32
756000 7c69ff 709ff DECW$SERVER_DDX_CFB
7c8000 8287ff 607ff DECW$SERVER_DDX_MFB

Exception Call stack dump follows:

PC IMAGE+offset of call
-- --------------------
119bd0 DECW$SERVER_DIX + e7bd0
104324 DECW$SERVER_DIX + d2324

********** marking the end of call stack dump **********
********************************************************

[Volker Halle]

Rick,

SYS$LIBRARY:DECW$SERVER_DDX_GY.EXE is included in VMS731_GRAPHICS-V0400. ANAL/IMA/INT should give a link date of 24-MAR-2005 - do you have the right image on your system ?

SYS$LIBRARY:DECW$SVEXT_MULTI_BUFFERING.EXE is not included in any OpenVMS patch, so it should be coming from the VMS V7.3-1 kit.

The secondary error messages are not correctly reported in the .LOG file, so there is no way to tell, WHY image activation failed.

You can still try ANAL/IMAGE on both files, to see if there would be any errors with those images.

Volker.

[Rick Dyson]

This system is remote and I can't get to it for some reason right now, but I know I did check the *DDY* exe and it was dated recently. I will check for those exact times later.

There were some recent power problems at the site that may have taken out some network component. I am told the system is still a "white screen". :)

When I applied GRAPHICS v4, I noted many files were not going to be overwritten due to "lower generation number". I have considered deleting at least the SERVER_DDX_GY exe (or others?) and reinstalling the ECO to see if it would replace the files. It could be that the files are corrupt?

Specifically:

%PCSI-I-RETAIN, file [SYSLIB]DECW$DRM_PRIV.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]DECW$DRM_RADEON.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]DECW$OPENGLSHR_RADEON.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]DECW$OPENGLUSHR_V11.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]DECW$OPENGLUTSHR.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]DECW$SERVER_DDX_FB.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]DECW$SERVER_DDX_RADEON.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]DECW$SERVER_DDX_XAA.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]DECW$SERVER_DIX.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]DECW$SERVER_DRI.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]DECW$SVEXT_GLX_RADEON.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSLIB]SDARMS$SHARE.EXE was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSMGR]DECW$DEVICE_CONFIG_GH.COM was not replaced because file from kit has lower generation number
%PCSI-I-RETAIN, file [SYSMGR]DECW$STARTSERVER.COM was not replaced because file from kit has lower generation number

But wouldn't Analyze/Image report something like that if it couldn't read the file? When I ran it yesterday on SERVER_DDX_GY, I know it didn't report any problems that I saw (though there were a lot of output pages!)

rick

[Benjamin Levy]

I had pretty much this same issue. I installed a brand-new DS-15 with a Radeon-7500 controller with V7.3-1, UPDATE-400, and GRAPHICS-300. I got the login banner on the graphics screen with no problem, but if I logged in the display would then become garbled and unreadable.

At HP's suggestion we upgraded to GRAPHICS-400, and it is much worse than before - I no longer get even the login screen in readable format - similar to yours with just white screen. The messages in the DECW server log file are similar to the ones you quote above.

As an experiment I then tried V7.3-2 with GRAPHICS-200 (comparable to GRAPHICS-300 on V7.3-1) and all is working well. I am currently investigating if I can stay at this level, or if I need to go back to V7.3-1 for other reasons and get this decwindows problem fixed.

There is one minor configuration difference between our and yours which I can't discuss in a public forum, but which is not likely to be the root cause explanation for this issue.

[Rick Dyson]

Exellent info Ben. I really have no problem going to v7.3-2. It was on my TODO list. The only real question for me was "why not try v8.2?" :) This is a developement/test box. So I have not broken it beyond acceptability. Plus, DECW is not used that much. Most connects are SSH or Telnet anyway.

Maybe I will still stick with v7.3-2 and save the v8.2 experimentation to a more "local, hands-on" machine!

Did you simply uprade to v7.3-2 and let it overwrite? Or did you need to cleanout or delete first? How about Motif? Did you upgrade? I believe I am running v1.3, but can't get the exact patch level since the box is not reachable yet...

Rick

[Volker Halle]

Rick,

you could try to directly run those 2 images with $ RUN SYS$LIBRARY:xxx.EXE

Maybe that way, we can get the real system status message causing the %LIB-E-ACTIMAGE

Consider to add a SET WATCH FILE/CLASS=MAJ command before the RUN (needs CMKRNL priv). It will tell you all other images accessed. Turn it off with SET WATCH FILE/CLASS=NOMAJ

Volker.

[Lawrence Czlapinski]

Rick, it's not likely but you might check your free global sections. If you were close to your maximum global sections, maybe the reboot pushed you over.
$FREE_GLBSECTS=F$GETSYI("FREE_GBLSECTS")
$SH SYM FREE_GLBSECTS
If you don't have enough free global sections at startup, DECwindows notes say it won't start up properly.
Lawrence

[Rick Dyson]

I have seen that happen many times, but you usually get a nice error message when you are booting on the console about needing to run AutoGen. Which never actually seemed to work the way it was designed, if memory serves. I had to always do it manually.

I even did that this time (just a general AutoGen with Feedback) because that was the first thing I assumed. I had not even gone looking for (or found) the DECW error log.

I can check specifically for that too when I can get to the box.

Boy, all these great suggestions at troubleshooting and I can't feed the info back to the forum! I'm really bummed! I just hope you all will still be here when I finally get the answer to all the questions!!

rick

[Rick Dyson]

A failed UPS kept me out, but now I have some results:

After a Set Watch...
$ run sys$share:DECW$SVEXT_MULTI_BUFFERING.exe
%XQP, Thread #0, Access (0,0,0) Status: 00000910
%XQP, Thread #0, Access DECW$SVEXT_MULTI_BUFFERING.EXE;1 (20322,2,0) Status: 00000001
%XQP, Thread #0, Control function (20322,2,0) Status: 00000001
%XQP, Thread #0, Access (0,0,0) Status: 00000910
%XQP, Thread #0, Access DECW$SERVER_DIX.EXE;1 (26484,2,0) Status: 00000001
%XQP, Thread #0, Control function (26484,2,0) Status: 00000001
%IMGACT-F-SYMVECMIS, shareable image's symbol vector table mismatch
-IMGACT-F-FIXUPERR, error when DECW$SVEXT_MULTI_BUFFERING referenced DECW$SERVER_DIX
%XQP, Thread #0, Deaccess (20322,2,0) Reads: 4, Writes: 0, Status: 00000001
%XQP, Thread #0, Deaccess (26484,2,0) Reads: 12, Writes: 0, Status: 00000001


Piping some Analyze/Image through Search for Sys$Share:DECW$SVEXT_MULTI_BUFFERING

image file identification: "DW V7.3-021217"
image file build identification: "X9J9-0060035000"
linker identification: "A11-50"

Same for Sys$Library:DECW$SERVER_DDX_GY:

image file identification: "DW V7.3-050324"
image file build identification: "PCSI-0060030000"
linker identification: "A11-50"

$ FREE_GLBSECTS=F$GETSYI("FREE_GBLSECTS")
sys_WOLF => Show Symbol FREE_GLBSECTS
FREE_GLBSECTS = 174 Hex = 000000AE Octal = 00000000256

[Volker Halle · Accepted solution]

Rick,

good, now you KNOW, why activating DECW$SVEXT_MULTI_BUFFERING failed !

There seems to be an inconsistency between DECW$SERVER_DIX and DECW$SVEXT_MULTI_BUFFERING. DECW$SVEXT_MULTI_BUFFERING on your system was apparently linked against an 'older' DECW$SERVER_DIX. DECW$SERVER_DIX should have been replaced by VMS731_GRAPHICS-V0400, but it failed (see your %PCSI-I-RETAIN error message !). If you want, try extracting DECW$SERVER_DIX from VMS731_GRAPHICS-V0400...

(see HELP/MESS SYMVECMIS for more)

Could you please do the same (RUN ...) for DECW$SERVER_DDX_GY ?

Then please forward this 'supporting' information to HP support to let them figure out, what went wrong.

Volker.

[Rick Dyson]

Late followup. I have run this through HP Support and they suggested essentially the same tests as was here. Their final recommendatation is to upgrade OpenVMS to v7.3-2, apply the ECOs out there and then if there is still a problem, re-open the call.

I hate that kind of answer, but in this case since it is mostly remote and I wanted to go to v7.3-2 anyway, I am planning on following that advice.

Unless someone else has any good final suggestions (except copying the contents of the install CD into Sys$System and Sys$Library like was done in the previous, related thread), I will just move on and close this thread.

I really appreciate all the help and suggestions up to now!!!!!

rick

[Galen Tackett]

Benjamin (and Rick, when you go to V7.3-2)

If you install VMS732_GRAPHICS V3, watch out when or if you install VMS732_UPDATE V4!

In the case of the Radeon 7500, at least, I found that VMS732_UPDATE V4 introduced a regression from VMS732_GRAPHICS V3, bringing back a 7500-specific problem I was having before I installed that graphics ECO.

For more complete information look at this topic:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=876649

[Galen Tackett]

I'm posting to this older thread because I'm now running into this problem too.

I just updated an Alpha DS10L V7.3-1 system to DECwindows V1.3 and the current (as of Jun 16) set of Alpha ECO kits, including the Graphics V4 mentioned here.

Back on July 12, 2005 (a couple of months after the most recent previous post to this thread) John Santos reported over on comp.os.vms that he'd run into this and solved it by restoring the original version of DECW$SERVER_DIX.

Do any of the HP folks (or other gurus) know what's going wrong when Graphics V4 is installed, and whether John's fix is the best way to get around it?

[Rick Dyson]

Opps, forgot to formally close this thread...

I did get resolution by upgrading to v7.3-2, but that was not really surprising. I did not test any further on the subject at v7.3-1, nor was I aware of the others reporting a similar problem that Galen mentioned! That at least helps my ego knowing it was maybe not something I had done! :)

Galen: sorry I can't be more help but maybe bringing this subject back up on the active list will help grab someone who can shed some more insight!

rick