Conference azur::mcc

    
    
    Just to complete the loop,
    
    	the process that has locked the tmp file is running the 
    
    [VMS$COMMON.SYSEXE]MCC_EXPORTER_FM_WRITER.EXE;2
    
    
    	This process is one of the exporter subprocesses owned by the
    background exporter batch job, that is still running on the batch
    queue.
    
    
    	The total number of files  in use by this subproces is as follows;
    
    PRICEC_3        0000047A  [VMS$COMMON.SYSLIB]CMA$LIB_SHR.EXE;3
    PRICEC_3        0000047A  [VMS$COMMON.SYSLIB]CMA$RTL.EXE;3
    PRICEC_3        0000047A  [VMS$COMMON.SYSLIB]CMA$OPEN_RTL.EXE;3
    PRICEC_3        0000047A  [VMS$COMMON.SYSLIB]CMA$OPEN_LIB_SHR.EXE;2
    PRICEC_3        0000047A  [VMS$COMMON.SYSEXE]MCC_EXPORTER_FM_WRITER.EXE;2
    PRICEC_3        0000047A  [VMS$COMMON.SYSLIB]SQL$INT.EXE;2
    PRICEC_3        0000047A  [PRICEC]MCC_EXP_SYNC_47A.TMP;1
    PRICEC_3        0000047A  [VMS$COMMON.SYSLIB]SQL$SHR.EXE;1
    PRICEC_3        0000047A  [PRICEC.MCC_RDB]CISCO_SNMP_EXPORT.RDB;1
    PRICEC_3        0000047A  [RDM$RUJ]CISCO_SNMP_EXPORT$0098A984CBAB7697.RUJ;1
    PRICEC_3        0000047A  [PRICEC.MCC_RDB]CISCO_SNMP_EXPORT.SNP;1
    
    
    	Could somebody give me a pointer as to which of these files might
    contain the reason as to why the exporter has "gotten lost"
    
    
    Regards, Conrad

Hi,

Some problems were reported about exporting of SNMP INTERFACE entities.

We did not manage to reproduce the problem here in VBO.

Could you add more details on how many interfaces are exported,
all interfaces on the same CISCO router or not, the polling period,
what is the elapsed time to get poll failures.

Thanks,

Benoit

    Hi Benoit,
    
    
    	Here's the command procedure I used to start the exporting;
    
    $! this routine should just start a background exporter job for snmp
    exports
    $! from Cisco1 and Cisco2
    $!
    $ set proc/priv=all
    $ submit/que=conrad_exporter/user=pricec/restart -
            /param=(uvonms$dka0:[pricec.mcc_rdb]cisco_snmp_export.rdb) -
                    sys$manager:mcc_exporter_background.com
    $!
    $!
    $ wait 00:01:00
    $!
    $ manage/enter
    !
    export snmp cisco2 interface 1 export target -
      uvonms$dka0:[pricec.mcc_rdb]cisco_snmp_export.rdb, export period
    00:30:00
    !
    export snmp cisco2 interface 2 export target -
      uvonms$dka0:[pricec.mcc_rdb]cisco_snmp_export.rdb, export period
    00:30:00
    !
    export snmp cisco2 interface 2 export target -
      uvonms$dka0:[pricec.mcc_rdb]cisco_snmp_export.rdb, export period
    00:30:00
    !
    export snmp cisco1 interface 1 export target -
      uvonms$dka0:[pricec.mcc_rdb]cisco_snmp_export.rdb, export period
    00:30:00
    !
    export snmp cisco1 interface 2 export target -
      uvonms$dka0:[pricec.mcc_rdb]cisco_snmp_export.rdb, export period
    00:30:00
    !
    export snmp cisco1 interface 3 export target -
      uvonms$dka0:[pricec.mcc_rdb]cisco_snmp_export.rdb, export period
    00:30:00
    !
            exit
    $ exit
    
    
    
    
    As you can see it is for two Ciscos, each is an AGS+ running V9.1(9)
    code.
    
    I received 10 succesfull polls, 28 failures (total 38 polls) , last
    failure for each of the entities being exported was an exception code 2
    or exception code 3.  I could understand the exception codes as the
    nearest cisco to the MCC station had it's ethernet interface
    interrupted for a few minutes, and this might have coincided with a
    poll attempt , but I wasn't watching closely enough to see if this was
    the case.
    
    However, the interface was put back on the cisco, and IP communication
    was available.
    
    
    
    		Regards, Conrad

Hi Conrad,

I am sorry but i cannot figure out what is your problem.

- As you remarked, files MCC_EXP_SYNC_nnn.TMP are opened by the process
  running the MCC_EXPORTER_FM_WRITER.EXE.
  This file is locked as sson as the process is running.
  MCC_EXP_SYNC_nnn.TMP files are not deleted as the process dies.
  This is a problem but not a major issue.

- What do you mean by " the exporter has gotten lost".
  If SNMP INTERFACE entities are unreachable and then are reachable again
  some polls are lost but polling may go on afterwards without entering
  any commands.

  What is the role of suspend export and delete export commands.
  Does the polling restarts correctly when entities are reachable again.

Regards, Benoit

    
    
    I'll check whether correct polling restarts, later on today , I'll do
    some tests.
    
    
    
    The "exporter getting lost" refers to the MCC replying that "there is
    no exporting for specified entity" , when several hours before hand it
    sad that "exporting has started", the .tmp files have been created, and
    no-one has issued a suspend or delete exporting command.  I decided to
    suspend then delete exporting ,as the ciscos I am experimenting with
    are very volatile and didn't want to upset the exporter process unduly,
    
    
    
    	I'll try out some more tests, and try to give a clear statement of
    what is happening.
    
    
    
    	Regards, Conrad