T.R | Title | User | Personal Name | Date | Lines |
---|
2888.1 | try this first | IOSG::STANDAGE | | Mon Jun 21 1993 09:31 | 18 |
|
Pete,
May I suggest you start up the FCS in the foreground to verify that
everything server specific is OK. At least we're then know it's
something directly related to startup procedures themselves...
To start the FCS in the foreground, do:
$ SRV =="$SYS$SYSTEM:OAFC$SERVER.EXE"
$ SRV OA$DATA_SHARE:<node>$SERVER73.DAT
Thanks,
Kevin.
|
2888.2 | fails | HGOVA::PETELAM | | Mon Jun 21 1993 09:57 | 22 |
| Hi Kevin,
I got the following result, I don't understand it, can you explain to
me ?
$ SRV =="$SYS$SYSTEM:OAFC$SERVER.EXE"
$ srv oa$data_share:dekfs1$server73.dat
%SYSTEM-F-INSFARG, insufficient call arguments
%MCC-F-FRAMEWORK_INIT, framework initialization failure
-SYSTEM-E-INSFARG, insufficient call arguments
%TRACE-F-TRACEBACK, symbolic stack dump follows
module name routine name line rel PC
abs PC
MCC_KERNEL_INIT MCC_KERNEL_INITIALIZE 178 0000003B
00086747
000866DB
000866DB
Thanks a lot indeed.
Pete
|
2888.3 | Never seen this before... | IOSG::STANDAGE | | Mon Jun 21 1993 12:36 | 24 |
|
Pete,
This is why the startup is failing - you have a problem with MCC.
MCC is a package for implementing multi-threading. The images included
with the FCS are a sub-set of the full MCC as the whole product was too
large to include.
I have never seen this problem before, but obviously somthing did not
survive the reboot (or something has been changed since) and hence you
have this problem.
Are there any MCC_* logicals defined on the system ?
I'll go into 'think mode'... (could take a while :-)
Kevin.
|
2888.4 | Have you rebuilt it ? | IOSG::CHINNICK | gone walkabout | Mon Jun 21 1993 13:15 | 16 |
|
Looks like a version mismatch between components...
What version of OpenVMS are you running?
Have you re-build FCS and copied it to SYS$COMMON:[SYSEXE]?
[Use OA$BUILD_SHARE:OAFC$SERVER_LINK.COM]
Have you copied the rebuilt OAFC$MTS_PRIV_SHR.EXE to SYS$COMMON:[SYSLIB]?
[And re-started FCS using @SYS$STARTUP:OAFC$STARTUP]
Failing these obvious checks and unless Kevin can suggest anything
further, it's going to be difficult to track down.
Paul.
|
2888.5 | yes...and no ! | IOSG::STANDAGE | | Mon Jun 21 1993 15:56 | 18 |
|
Pete,
Paul is correct, if you've got a new version of OpenVMS, then ALL-IN-1
and the FCS should be relinked. If this is the case, the FCS will be
automaticically relinked when A1LNKDRV is run. However, you can indeed
link it seprarately via OAFC$SERVER_LINK.COM.
One correction to Paul's note, please do not start the FCS using
OAFC$STARTUP.COM. This is an unsupported (and I believe undocumented)
method of starting the FCS which will lead to problems. The FCS should
always be started by running A1V30START.COM, or via the Manage Servers
subsystem (SM MFC MS STA).
Kevin.
|
2888.6 | more info | HGOVA::PETELAM | | Mon Jun 21 1993 19:34 | 27 |
| Hi Kevin, Hi Paul,
Thanks a lot for your replies.
The VMS is v5.5-2/Hangul version. It worked at the very beginning
before they reboot their system.
I have checked the .exe files and they are in place, but they are
not installed (but I don't think reboot will remove the entries in
database, maybe I need to check the manual and see if any images
need to be installed tomorrow in the office).
After I started the server (SM MFC MS STA) I could see (by $ sh sys)
<host>$73srv, but it pass away very soon, very strange !
However, without that process I still can create new drawer and
switch around ...
I am still very puzzled with the "open/read/write" statements,
I think the "/read" is redundant in the file and will cause
warning message if you run the a1postinstal.com.
Thanks a lot again.
Cheers
Pete
1:30 am at home :-(
|
2888.7 | "Files" are really mail boxes | CHRLIE::HUSTON | | Mon Jun 21 1993 22:20 | 26 |
|
re .6
>I am still very puzzled with the "open/read/write" statements,
>I think the "/read" is redundant in the file and will cause
>warning message if you run the a1postinstal.com.
The "file" that is being opened is a mailbox. The FCS startup
uses VMS mailboxes to pass a couple of things back and forth.
The process, being there, then going away is consistent with most
FCS Startup failures, the FCS startup takes a few seconds and this
means that it got far enough to actually do the $run/detached from
the command file, but sometime later, something went wrong.
I would also recommend trying a re-link.
I saw this type of thing last week on one of our systems, I will
check with the system manager and see what it was. I didn't know off
the top of my head. They tried installing AIDA and that is what
triggered the problem.
I will check and get back to you if they figured out what happen.
--Bob
|
2888.8 | thanks | HGOVA::PETELAM | | Tue Jun 22 1993 06:03 | 9 |
| Thanks Bob,
But the re-link does not help. Since they can see the drawer and can
use WP and EM (I "fixed" it without actually know how I fixed it :-) ),
it will allow me to take some times to study how to restart the
fc-server.
Cheers
Pete
|
2888.9 | Did you re-start after re-link/copy | IOSG::CHINNICK | gone walkabout | Tue Jun 22 1993 10:49 | 16 |
|
Pete...
The /READ/WRITE is correct and is not relevent to this problem. Don't
change this from what is supplied!
The images DEFINITELY NEED TO BE INSTALLED. This is done for you when
you run A1V30START. If you don't run this after copying your re-linked
images into the appropriate directories, the new images wont take
effect. Check that when you run the startup that you aren't getting
failures during the image install section.
Paul.
|
2888.10 | do this, and let us know... | IOSG::STANDAGE | | Tue Jun 22 1993 12:03 | 29 |
|
Pete,
Here's what you should have done:
$ SET DEF OA$BUILD_SHARE
$ @OAFC$SERVER_LINK
This produces two images:
OAFC$SERVER.EXE
OAFC$MTS_PRIV_SHR.EXE.
Now...
$ COPY OAFC$SERVER.EXE SYS$COMMON:[SYSEXE] /log
$ COPY OAFC$MTS_PRIV_SHR.EXE SYS$COMMON:[SYSLIB] /log
** Check that no version of these files exist in the SYS$ROOT:
directories, otherwise they'll be picked up first and you'll get
mismatches in images.
Now run A1V30START.COM and see if your server is running...
Kevin.
|
2888.11 | I have tried to relink but ... | HGOVA::PETELAM | | Tue Jun 22 1993 17:35 | 17 |
| Kevin,
I have done that but the <host>$server73srv still pass away
within few second. Now I suspect their sysgen parameters have
changed since reboot, however, why there is no error message ?
Anyway, I will go through that tomorrow.
Could you give me some advice about MCC ? I have no idea what it
is and how to start it (I cannot find its startup file in sys$startup).
The only file is oafc$sysstartup.com (?) which is called by
oafc$startup.com for .exe installation.
Cheers
Pete
PS Now I realize why the system can locate the MAIN drawer, it is
because I have started the server for few seconds.
|
2888.12 | | IOSG::STANDAGE | | Tue Jun 22 1993 17:43 | 20 |
|
Pete,
Briefly (in the middle of something)...
Your MAIN drawer will always be visble no matter what, it is not
dependant on the FCS being started. The FCS comes into play when you
start performing cross drawer operations, etc.
MCC is packaged WITH the FCS, therefore you will not find any startup
procedures for it.
From within ALL-IN-1, try creating a new server (object 222 for
instance) and starting it in the forground - does this give the same
results ?
Kevin.
|
2888.13 | still has no clue ... | HGOVA::PETELAM | | Wed Jun 23 1993 11:42 | 9 |
| Kevin,
Do you think the <node>$server73.dat is corrupted ? How do I know it is
corrupted ?
I have checked the system parameters and they are alright.
Thanks
Pete
|
2888.14 | problem solved | HGOVA::PETELAM | | Wed Jul 07 1993 05:51 | 12 |
|
Hi,
The problem is solved by:
- disable net$startup in sysstartup_v5.com
- enable startnet in the same file
reboot.
Pete
|