Hello Paul This was a very good followup. If I might add two things.
1) We defined schedules that ran from 17:30 hrs and a duration of 12
hours with a short randomization time. The main purpose of setting
the schedule so long was to allow machines that were unavailable
for backup in the evening hours to start there backup if they
became available during the middle of the night (Could be that the
unit was down or network problems). Also if for any reason the
backup was halted (client or server problem) and the scheduler
connected again to the server, if the schedule window was still
open it would restart the backup.
2) We found that with a large number of clients in a server and with
a number of them unreachable for some reasons, the method used to
queue the prompted requests would lead to an underutilized server
and clients with missed schedules and no error messages. I believe
that a change in the way that adsm handles it's queuing would
fix this problem. (It is much like the queuing problems with
smtp).(Of course since IBM does not offer souce code or
internals doc, The exact nature of the problem is a guess).
But the short of it was that for us polling was much better at
getting the job done. But it did lead to many messages in the
server log about no sessions available. But with polling the cleint
trys again later.
I suspect in this case the problem is none of the above since there
is an error message.
Was a schedule deleted or renamed? Was the assoc between the node
and schedule deleted. Check again to see if the duration is 8hours.
Let us know what you find.
Thanks len boyle
-------------------------------------------------------------------------
Leonard Boyle snolen AT vm.sas DOT com
Leonard Boyle snolen AT vm.sas DOT com
SAS Institute Inc. ussas4hs@ibmmail
Room RB448 len.boyle AT sas DOT com
1 SAS Campus Drive (919) 677-8000 ext 6241
Cary NC 27513
----------------------------------------------------------------------
In article <002201bf1081$ffab3860$0a00a8c0 AT yohann.cgws1.ab.wave.home DOT
com>, Paul
In article <002201bf1081$ffab3860$0a00a8c0 AT yohann.cgws1.ab.wave.home DOT
com>, Paul
Fielding <paul.fielding AT HOME DOT COM> says:
>
>The short answer is, yes, they can span midnight. Very common.
>The long answer.....
>
><TUTORING CAP>
>
>2) If the node is set to SCHEDMODE POLLED (not preferred), when the node
>contacts the server at 11:05, the server will deny the request and say it's
>missed it's window.
>
>Note that if you don't have a SCHEDMODE line in your options file, it will
>default to POLLED mode.
>later,
>Paul
>
>
>-----Original Message-----
>From: Ryan Carter <racarter AT DATA-TRONICS DOT COM>
>To: ADSM-L AT VM.MARIST DOT EDU <ADSM-L AT VM.MARIST DOT EDU>
>Date: Wednesday, October 06, 1999 10:48 AM
>Subject: Can a client backup schedule span midnight?
>
>
>>I am running ADSM 3.1.22 on OS/390 2.6. I have a client backup schedule
>>that begins at 20:00 and is set to last 8 hours. I am finding that many
>>clients are not being backed up each night and a message is being
>>written to the log stating: "ANR2578I Schedule BACKUP21 in domain
>>DOMAINNAME for node NODENAME has missed its scheduled start up window."
>>Last night I watched a machine check in after midnight but before the
>>schedule finished. When it checked in, it established a session and
>>then generated the error "ANS1815E Either the window has elapsed or the
>>schedule has been deleted"
>>
>>Am I therefore to conclude that a backup schedule can not cross
>>midnight?
>>
>>Thanks,
>>Ryan Carter
|