> Date:         Fri, 18 Aug 1989 11:41:47 EDT
> From:         Gary Sponseller <SPONSELL@AKRONVM>
>
> Unfortunately for *all* Listservs netwide, the 400-odd individual
> SIGNOFF * ( NETWIDE commands, for userid@SUNRISE, issued on 8/16/89,
> caused quite a bit of local resource usage at each site!!  >8-(
> On Akronvm alone, the Listserv log increased in size by 1000% and
> Listserv CPU time consumption increased by 700%!  These are figures
> compared to usual values for local daily Listserv resource usage.  (Except
> for the SUNRISE deletes, no significant Listserv tasks were notable on 8/16.)
> Listserv sites with many lists would have used quite a lot more resources.
>
> As the NAD (Node Administrator) for SUNRISE, PLEASE create a Listserv
> job and send it to [log in to unmask]  The deletes will be done in a batch
> and performed as efficiently as possible.  I do this with class userids
> once a semester -- the number of userids is in the thousands.
 
    I have already appologized directly to Gary, but I would like to take this
opportuinity to appologize to all LISTSERVer adminstrators for the resource
drain I caused trying to SIGNOFF all the students no longer enrolled at the
University.  I was following the advice of another NAD and did not realize
there was a more efficient method for accomplishing this task.
 
    I can imagine the situation was not amusing to any of you, but I found the
following chain of events very amusing:
 
1)  I asked our "global" campus NAD to make this UserName NAD for Sunrise
2)  Upon confirmation of the changes from BITNIC, I tried an UNSUBSCRIBE for
someone else and was told I was not the NAD for SUNRISE
3)  So, I asked the "global" NAD to UNSUBSCRIBE the party and he was told he
was not the NAD for SUNRISE,
4)  Nor was he NAD for his own node any more -- It seems that rather than
just adding me, they had replaced him
5)  So, I sent the large number of SIGNOFFs
6)  Gary pointed out the error of my ways
7)  So, I tried the DELETE job and got the following responses:
 
> Date:         Thu, 24 Aug 89 19:50:25 LCL
> From:         Revised List Processor (1.6a) <LISTSERV@SUVM>
>
> Your request will be forwarded to the nearest backbone server, [log in to unmask]
 
> Date:         Thu, 24 Aug 89 19:53:11 EDT
> From:         Revised List Processor (1.6a) <LISTSERV@PSUVM>
>
> Your request will be forwarded to 187 servers.
 
    I received mail from 8 LISTSERVers saying they had successfully processed
my request, but 74 complained as follows:
 
> Date:         Thu, 24 Aug 89 19:54:33 EDT
> From:         Revised List Processor (1.6a) <LISTSERV@PSUVM>
>
> You are not a registered NAD (Node Administrator) for node SUNRISE.
> The NAD for node SUNRISE is [log in to unmask]
> LISTSERV management has been notified.
 
    I found it incredible that <ListServ@PSUVM> forwarded my NetWide Delete
request and then refused to honor it.  I seems to me it would have made more
sense for that LISTSERVer to just reject it in the beginning.
 
    Note, also, that many of the rejection messages included the following:
 
> A REXX error occured while processing your command.
> Part of the command output might have been lost.
 
    Again, my appologies for the problems I caused and thanks for allowing
me to share some of the irony of the situation.
 
Ed Lyon, PostMaster and NAD for SUNRISE
Syracuse University Computing and Network Services
 
P.S.  Are there still 105 LISTSERVers out there waiting to tell me I goofed?