Conference iosg::all-in-1_v30

3151.0. "OA$LIB:SM_FC_SERVER_STOP.SCP interfere" by UTRTSC::SMEETS (Martin, CSC Office support (NL)) Mon Aug 16 1993 16:12

Hi,

In order to be sure that all ALL-IN-1 (data)files are closed at the beginning of
a backup operation, we provided to a customer the following command procedure.

$! A1V30STOP.COM Stop procedure for ALL-IN-1 Version 3.0.
$!
$!      This procedure DEINSTALLes and STOPs all ALL-IN-1
$!      images and processes.
$!
$!      It should be run from a VMS privileged ALL-IN-1 account.
$ set noon
$ set noverify
$ write sys$output  ""
$ write sys$output  "           This procedure deinstalls ALL-IN-1 3.0 files,"
$ write sys$output  "           to prevent OPEN files on ALL-IN-1 disk(s)"
$ write sys$output  ""
$!
$ write sys$output  "           1) Remove installed form libraries and TXL's"
$!      Remove installed form libraries and TXL's
$allin1/noinit/override
OA$FLO_OPEN_LIB OA$LIB:MANAGER
DO OA$LIB:SM_FC_SERVER_STOP
DO OA$LIB:SM_FC_SERVER_STOP
oa$fbt_remove_library oa$lib:memres
oa$fbt_remove_library oa$lib:oaformoa$fbt_remove_library oa$lib:sitememres
oa$fbt_remove_library oa$lib:siteoaform
oa$txl_remove a1txl
oa$txl_remove cmtxl
exit
$ write sys$output  "           2) Remove installed images"
$!      Remove installed images
$install:=="$sys$system:install /command_mode"
$install
remove oa$lib:mailcount
remove oa$lib:oa$submit
remove oa$lib:oa$main
$!
$ write sys$output  "           3) Stop process OAFCV"
 context = ""
$STOP_OA$FCV:
$!
$ pid = f$pid(context)
$ if pid .eqs. "" then goto END_STOP_OA$FCV
$ name = f$getjpi(pid, "prcnam")
$ if name .nes. "OA$FCV" then goto STOP_OA$FCV
$ stop /id='pid'
$
$END_STOP_OA$FCV:$!
$ write sys$output  "           4) Stop process TM_SERVER"
$ context = ""
$STOP_TM_SERVER:
$!
$ pid = f$pid(context)
$ if pid .eqs. "" then goto END_STOP_TM_SERVER
$ name = f$getjpi(pid, "prcnam")
$ if name .nes. "TM_SERVER" then goto STOP_TM_SERVER
$ stop /id='pid'
$
$END_STOP_TM_SERVER:
$!
$! Deassign ALL-IN-1 main image logical: sender and fetcher will wait...
$ deass /system /exec oa$image
$! Deassign oa$lib_share (checked in A1V30START.COM).
$ deass /system /exec oa$lib_share
$!
$ write sys$output  ""
$ write sys$output  "           Deinstall ALL-IN-1 3.0 finished"
$ write sys$output  "           Run sys$manager:a1v30start.com to restart"
$ set verify
$ EXIT

This procedure which runs on BOTH nodes (LIM003 and LIM004) of a cluster has 
ran for several months (7 days a week) without any problems. Last week the 
customer found in the logfile of node LIM004 the following error lines:

 This procedure deinstalls ALL-IN-1 V3.0 files,
 to prevent OPEN files on ALL-IN-1 disk(s)
 1) Remove installed form libraries and TXL's
Stopping server LIM003::"73="
Stopping server LIM004::"73="
OAFC Error Report : %OAFC-E-INTERR, Internal error in File Cabinet Server

The server logfile cointains the following error lines:

11-AUG-1993 20:15:17.83  Server: LIM004::"73="  Error:
%MCC-E-ALERT_TERMREQ,
 thread termination requested  Message:
SrvTimeoutSysMan; receive alert to terminate thread

11-AUG-1993 20:16:13.83  Server: LIM004::"73="  Error:
%MCC-E-EXISTENCE_ERROR,  object does not exist
 
Could it be that the two SM_FC_SERVER_STOP scripts by accident interfere with
each other ? 

Or any other thoughts on what's going on here.

Thanks,

Martin


p.s Since 12 august the error didn't show up anymore

    
    
    FYI,
    
    One of your errors is in STARS article :
    
    Expected Error Messages In File Cabinet Server Log When Shutdown 
    
    (I searched for MCC-E-ALERT_TERMREQ,)
    
    Message 2:                                                                    
    ___________________________________________________________________________   
    30-JAN-1992 15:32:18.76  Server: SMOOSE::"73="
    Error:%MCC-E-ALERT_TERMREQ,    
    thread termination requested  Message: SrvTimeoutSysMan;                      
    receive alert to terminate thread                                             
    ___________________________________________________________________________   
                                                                                  
    Explanation:                                                                  
                                                                                  
    This indicates that a background thread received a signal to abort, as        
    this signal was not expected it is recorded as an error, however the          
    message is really informational and can be ignored.
    
    Clive

Hi Clive,

I was aware of this specific STARS article, but anyway thanks.

The problem at the customer's site however was that the procedure 
OA$LIB:SM_FC_SERVER_STOP didn't continue after it reported the internal error.

Regards,

Martin

Hi to all of you,

I guess my previous description of the problem (.0) wasn't clear enough, so I
will give it another try.

	+--------------+	+---------------+
	|              |	|               |
	| NODE_A$SRV73 |	| NODE_B$SRV73  |
	|              |	|               |
	+--------------+	+---------------+
        

On both cluster nodes (NODE_A & NODE_B) the filecabinet server is running. 
Now I run the procedure OA$LIB:SM_FC_STOP_SERVER.

What are the (known) problems one could expect when running 
OA$LIB:SM_FC_STOP_SERVER at the same time of both nodes.

The customer is pushing me for a solution. Is it possible to avoid these/this
problem ?

Thanks,

Martin