| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 3101.1 | | METSYS::THOMPSON | | Wed Apr 16 1997 11:22 | 7 |
| Hi,
this is the second report of this. It didn't happen on our systems.
We need to investigate.
M
|
| 3101.2 | | METSYS::THOMPSON | | Wed Apr 16 1997 11:50 | 5 |
|
My first reaction would be that this is an ObjectBroker bug, did
you report it to them?
Mark
|
| 3101.4 | decedi$configure_acas or obb bug? | CSC32::B_ALTMIRE | | Wed Apr 16 1997 14:57 | 32 |
| Mark,
I was involved in the call referenced by Lorey in note 3080. I was
never able to get dialed in to the customers system but he reported the
same errors:
....
Configuring DEC ObjectBroker...
%OBB-E-DB_OBJNOTFND, Object `ROOT' not found.
$ @decedi$tools:decedi$configure_acas
....
Configuring DEC ObjectBroker...
%OBB-E-DB_OBJNOTFND, Object `1$Definitions' not found.
I did get OBB involved and at their request deleted all OBB images and
files and reinstalled but got the same results when running
decedi$configure_acas. The final resolution was when Obb dialed in and
reinstalled OBB and got essentially the same results at first. The
final resolution was after all implementations had been trashed, including
the OBB_SOFTWARE_REGISTRY implementation. At that point he deleted all
copies of OBB_SOFTWARE_REGISTRY.REG from obb$library and reregistered
the OBB_SOFTWARE_REGISTRY with:
$ appl/broker import registry /implementation obb_soft_reg.dat
At that point he was able to run decedi$configure_acas sucessfully and
start DECEDI. Unfortunately I don't know whether DECEDI was up or down
doing most of this or if it mattered.
By the way there is no note 2986.6 you referenced in 3080.2.
Bill
|
| 3101.5 | | METSYS::THOMPSON | | Wed Apr 16 1997 18:32 | 5 |
| Bill,
so are ObjectBroker acknowledging it is their bug?
Mark
|
| 3101.6 | I can reproduce easily | RULLE::KLASSON | Sven-Olof Klasson @GOO | Wed Apr 16 1997 19:21 | 33 |
| Hi,
re: .1
> this is the second report of this. It didn't happen on our systems.
I get exactly the same error messages as the customer.
Try this to reproduce:
Delete the following files
SYS$SPECIFIC:[OBB...]*.*;*
SYS$SPECIFIC:[000000]OBB.DIR;1
SYS$COMMON:[OBB...]*.*;*
SYS$COMMON:[000000]OBB.DIR;1
SYS$MANAGER:OBB*.*;*
SYS$COMMON:[SYSEXE]DECEDI_OBB_CONTEXT.CO;*
SYS$COMMON:[SYSEXE]OBB$*.EXE;*
SYS$COMMON:[SYSLIB]OBB*.*;*
Install Objectbroker V2.7-10 (this is the kit on the CD's)
@DECEDI$CONFIGURE_ACAS
It seems like if OBB is upgraded from V2.6 to V2.7 DECEDI$CONFIGURE_ACAS will
run without errors.
Mark, if you want, you can logon to my system and have a look. I only
use DEC/EDI and Objectbroker for testing, so you can do what ever test you like.
Send me a mail if you want to do this, and I'll give you access.
re: .2
> My first reaction would be that this is an ObjectBroker bug, did
> you report it to them?
No, I have not reported this to ObjectBroker engineering.
/Sven-Olof
|
| 3101.7 | | MUNICH::WWERNER | When in doubt, do as the INTs do | Thu Apr 17 1997 10:32 | 4 |
| I have already reported the problem in OBJECTBROKER_DEVELOPMENT note 2466.
Not much reaction up to now. Maybe an IPMT is neccessary...
Wolfgang
|
| 3101.8 | | METSYS::THOMPSON | | Thu Apr 17 1997 12:36 | 8 |
|
All DECEDI$CONFIGURE_ACAS is is dcl script. There is no DEC/EDI code
running in this, all that we supply is the Object References and IDL,MML etc..
This probably should be IPMT's asap.
Mark
|
| 3101.9 | Workaround | RULLE::KLASSON | Sven-Olof Klasson @GOO | Mon Apr 28 1997 10:19 | 26 |
| Hi,
The following workaround seems to work. I have used this on my test system.
Try this if you have done a fresh installation of Objectbroker V2.7.
If you have uppgraded Objectbroker to V2.7 from a earlier version I don't
think you'll see this problem.
1) If DECEDI$CONFIGURE_ACAS.COM has been run and failed, it has left
corrupted SYS$COMMON:[OBB]OBB_SOFTWARE_REGISTRY.REG files on the system.
These files are not automatically purged.
Delete the corrupted files. But leave the older versions.
(A fresh installation of OBB V2.7 puts two versions of this file on the
system).
2) Issue this command
$ APPL/BROK REG IMPL/REPLACE DECEDI$TOOLS:DECEDI$IMR_REG.DAT
3) $ @DECEDI$TOOLS:DECEDI$CONFIGURE_ACAS
/Sven-Olof
Ps There is a IPMT on this case opened by MUNICH::WWERNER
The IPMT case number is CFS.50543 / MGO102787 (prio 3)
|
| 3101.10 | | METSYS::THOMPSON | | Mon Apr 28 1997 11:22 | 7 |
|
Ah ha! That's why it's not getting much airtime.
This is a "system down" situation until they get your
workaround. I would have put that at P1.
Mark
|
| 3101.11 | DECEDI$CONFIGURE_OBB.COM | METSYS::THOMPSON | | Wed May 07 1997 16:22 | 28 |
|
In EDIENG""::user2:[decedi_v21d_patches] there is a new
procedure: DECEDI$CONFIGURE_OBB.COM.
For V2.1D systems, place this in DECEDI$TOOLS: and it can be used
as a replacement for DECEDI$CONFIGURE_ACAS.COM.
It incorporates a "workaround" from BEA Systems Inc.. This procedure
should *avoid* the problems reported in .0. If the User has already
been affected by this then contact the CSC for details as to how to
recover.
Note: this only seems to be an OpenVMS Alpha problem. BEA Systems have
not determined the casue yet but it appears that the changes in
DECEDI$CONFIGURE_OBB.COM will avoid the problem.
We have never seen this on our internal systems (in Reading). It has been
stated that if a Clean install is performed for V2.7 then problems will
result. However we have a least one customer who did a clean install
and found no problems.
It is possible this problem is quota/sysgen related, which may explain why
some see it and others do not.
Mark
|