[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference iosg::all-in-1_v30

Title:*OLD* ALL-IN-1 (tm) Support Conference
Notice:Closed - See Note 4331.l to move to IOSG::ALL-IN-1
Moderator:IOSG::PYE
Created:Thu Jan 30 1992
Last Modified:Tue Jan 23 1996
Last Successful Update:Fri Jun 06 1997
Number of topics:4343
Total number of notes:18308

1065.0. "File cabinet server fails to startup." by WAYOUT::CROOKS (Things that make you go Hmmmmm....) Thu Jul 16 1992 19:51

Hi folks,

I have looked through all the other notes in here along these lines and drawn
a blank so....

Customer cannot get their FCS working....batch job fires off no problem, they 
get the process created ok and that hangs around but they do not get the 
Decnet object and the status never changes from Stopped. (A1 startup gets TM 
server + obj_72 going alright so doesn't look like problem creating Decnet 
objects.) They have not had the FCS going yet since the Installation.

In the Installation log they got 

%OA-I-LASTLINE, "ALL-IN-1 running -- Starting local File Cabinet Server"
%RUN-S-PROC_ID, identification of created process is 20E048A5
%OA-I-LASTLINE, ERROR STARTING SERVER SUN::"73="
%OA-I-LASTLINE, "ALL-IN-1 running -- Populating partition data"
%OA-I-LASTLINE, File Cabinet Server SUN::"73=" is not available


However now if they do STA from the menu the log looks like a perfect startup
but they do not have obj_73 and the status is Stopped. 

If they start the server interactively from $ they get the message

15-JUL-1992 14:08:04.12  Server: SUN::"73="
 Message: Startup for File Cabinet Server V1.0 complete

But its not, so where do I go from here....? 
(sorry Terry that blows point 4 of Diagnosing Startup Problems in the 
FCS Bible....:-)

Thanks in advance Alan.

(They are only running ALL-IN-1 on 1 of 3 nodes in the cluster but the batch
queue is not generic and is ON the ALL-IN-1 node)
T.RTitleUserPersonal
Name
DateLines
1065.1Questions ...YR2000::VOLLER_IGordon (T) Gopher for PresidentThu Jul 16 1992 21:4311
    Alan,
    
    	Anything in the server logs (SYS$MANAGER:OAFC$SERVER or
    	OAFC$SERVER_ERROR)?
    
    	What completion status does accounting have for the batch job and
    	the detached (*SRV73) jobs?
    
    Cheers,
    
    Iain.
1065.2Checked all the previous ones?YR2000::WICKS_ADEC Mail Works for ME sometimesThu Jul 16 1992 21:4611
    Alan,
    
    What's in the log files? have you tried all the STARS articles about
    PROFILE.FDL, EPM$SRVSHR, SCSNODE etc ... - way too many to post.
    
    Also what if you put SET VERIFY on and run OAFC$STARTUP.COM?
    
    Regards,
    
    Andrew.D.Wicks
                                                                
1065.4WAYOUT::CROOKSThings that make you go Hmmmmm....Fri Jul 17 1992 12:1317
Hi guys,

As I say the log files tell ya nothing. OA$LOG:OAFC$SERVER_STARTUP.LOG looks
normal - it does the preliminaries starts the process, opens the mailboxes etc
they see "All done so go home!". There is next to nothing in the sys$manager
logs.

The thing is every other article or note mentions an error somewhere - they
have none.
 
Anyway I'll check the exit status. Set verify + oafc$starup.com will look
exactly like the oa$log startup log file wont it?

Thanks so far....

Alan.

1065.5something else to tryIOSG::STANDAGEOink...Oink...MoooooooooooooooooooooooooooooooooSun Jul 19 1992 22:1113
    
    
    You could also just try running the server in the foreground and seeing
    if it complains then :
    
    $ SRV=="$SYS$SYSTEM:OAFC$SERVER.EXE"
    $ SRV OA$DATA_SHARE:<node>$SERVER73.DAT
    
    
    
    Kevin.
    
    
1065.6WAYOUT::CROOKSThings that make you go Hmmmmm....Mon Jul 20 1992 18:5711
re .5 thats in .0 - it gives the successful completion message.

re .3 haven't got an exit status because the process doesn't die unless
they stop/id it.

This is going to get a bit hot if I dont get this working soon so
any flashes of inspiration welcome to get some idea of whats 
occurring.

thanks Alan

1065.7A few things to do...CHRLIE::HUSTONMon Jul 20 1992 20:0223
    
    My first guess, given what everyone has said, and the fact that
    starting in the foreground is ok, is that the config file that is
    used by the STA option and the config file used when starting
    in the foreground are not the same, and the one used by the STA option
    is using a non-73 object number.
    
    this would explain several things, like you say it looks as though it
    starts, but no srv73 shows up.
    
    If you can get to the point were it seems that the FCS is up, do a 
    $ sho dev/files/nosys and see if you can find a FCS process and see
    which config file it has open.
    
    Also can you post the log files, you say they have next to nothing, but
    maybe they have more than you think. To avoid alot of stuff in them
    can you delete the old versions (or rename them) and then start the
    servers.
    
    Also, by any chance is SUN the cluster alias??
    
    --Bob
    
1065.8More info.WAYOUT::CROOKSThings that make you go Hmmmmm....Fri Jul 24 1992 18:5311
Hi Bob thanks for the response,

The log file is on WAYOUT::OAFC$SERVER.LOG ( for the sake of fellow Windows
noters. )

When they run the image interactively they get the process started as usual but
a sh dev/files reveals that the process does not have any files open! 

SUN is the node name not the cluster alias.

cheers Alan. 
1065.9Delete and recreateIOSG::TALLETTArranging bits for a living...Fri Jul 24 1992 20:4311
    
    	Try deleting the server entry from the server master and
    	re-creating it. Don't forget to write down all the field
    	values before you delete, but you can accept most of the
    	defaults that Create offers you.
    
    	This will recreate the server configuration file and seems
    	to cure a few ills.
    
    Regards,
    Paul
1065.10Sorry bout that...CHRLIE::HUSTONFri Jul 24 1992 20:5518
    
    re .8
    
    Sorry, you need to do the $ sho dev/files/nosys on the disk that
    holds oa$data_share. If it still shows up that no FCS process has
    any files open then you have a problem, the following files at least
    should be open:
    
    partition_master.dat
    profile.dat
    partition.dat
    some sort of configuration file
    oa$script_completion.dat
    
    If there are not files then there is no server.
    
    --Bob
    
1065.11This is beginning to get to me......:-(WAYOUT::CROOKSThings that make you go Hmmmmm....Wed Jul 29 1992 14:3731
Hi Paul, Bob

re .9 we've tried that I'm afraid.

re .10 sorry we assumed that we were going to get a process SUNSRV73 and were
looking for files opened by that process as opposed to the one running the .exe,
which is just the ALLIN1 account.

Anyway it has all the expected files open including oa$data:sun$server73.dat. 

I'm sorry but I dont see how STA could be looking at a different config file, 
but then I dont see a lot sometimes...:-). Anyway I've got them to Edit the 
SUN::73= record in SM MFC MS and the config file specified there is the same one
used to run the image interactively.

The thing that I cant understand is why the OBJ_73 is never created in NCP.
When exactly is this done I cant find anything about it?

thanks for the continued interest

Alan.







used to run image interactively.

The thing that I cant understand is why the OB
1065.12Send me your config file...CHRLIE::HUSTONWed Jul 29 1992 15:0414
    
    re .11
    
    Ok this is getting bad, can you mail me the config file that you are 
    using to start the server? Send it to CHRLIE::HUSTON, i will see if
    I can find anything wrong with it.
    
    The decnet object is created fairly early in server life so if the
    server is up the decnet object is there. When you believe the server to
    be up and you can see what files it has open, is the process that is
    running the .exe show up in a mcr ncp sho know objects command??
    
    --Bob
    
1065.13Oooops....meant to write this B4 I went on holiday...WAYOUT::CROOKSThings that make you go Hmmmmm....Mon Aug 10 1992 13:309
Sorry Bob just as you were becoming intrigued.....the customer has
given up and prompted by a couple of other problems has decided that
they are going to restore v2.4 and do another upgrade from scratch.
If they have any problems after that they will be back.....but for
now fingers crossed.

thanks to everyone....

Alan.