When a scheduled client operation completes, the scheduler will notify the
server that the operation either completed successfully (QUERY EVENT
status shows "Complete") or failed (QUERY EVENT status shows "Failed").
The scheduler will also provide a result code, which you can see via QUERY
EVENT with the FORMAT=DETAILED option. The client scheduler does not
report any other status, so if you are seeing a status of "Missed" for a
particular event, then the client event never started to begin with. Some
reasons this could occur include:
- If prompted mode is being used, the server was unable to contact the
client to start the scheduled event.
- The scheduler wasn't running for the duration of the schedule window.
- The scheduler was running, but was busy with some other scheduled event.
An extreme situation could be the case where Monday night's backup runs so
long, it is still running during Tuesday night's backup window (i.e.
backup is taking more than 24 hours to run).
In the case you are talking about, I would suggest reviewing the following
items to see if you can piece together what is going on:
- Client dsmsched.log, dsmerror.log, dsierror.log, and TDP logs (if this
is a TDP backup scheduled via the client scheduler).
- Server activity log for the past few days, looking for messages related
to the node name.
- QUERY EVENT FORMAT=DETAILED for the past few days to see the status of
previous events.
- Server accounting records for the past few days to see information about
the node (what it did, and when).
Some questions that you should ask yourself (not necessarily an exhaustive
list):
1) Look at the session(s) for this node that are currently running. When
did they start?
2) Did the sessions start during the schedule windows?
3) Are there any messages regarding broken communications (i.e. TCP/IP
connection failures)?
3) Are there any messages in the activity log that show the server trying
to prompt the client (if using prompted scheduling mode)?
4) What does QUERY EVENT show for the past few nights' events for this
node?
5) What activity is in the dsmsched.log file regarding event start and
end? For what events, and when? (Look at timestamps carefully.)
6) Are there any error or warning messages in dsmsched.log, dsmerror.log,
dsierror.log, and/or TDP (if applicable) logs?
Regards,
Andy
Andy Raibeck
IBM Software Group
Tivoli Storage Manager Client Development
Internal Notes e-mail: Andrew Raibeck/Tucson/IBM@IBMUS
Internet e-mail: storman AT us.eyebm DOT com (change eye to i to reply)
The only dumb question is the one that goes unasked.
The command line is your friend.
"Good enough" is the enemy of excellence.
Gene Walters <gwalters AT GWMAIL.STATE.WV DOT US>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
04/21/2003 06:21
Please respond to "ADSM: Dist Stor Manager"
To: ADSM-L AT VM.MARIST DOT EDU
cc:
Subject: Re: Query Event VS Successful Flag
What determines how the event status gets set? I have another server,
that gave me a status of "missed", however, when I look at currently
active sessions, I can see the server still backing up, so it wasn't
missed. It didn't finish during the scheduled backup window, but I'm
having a hard time understanding what detrmines how the event status
gets set.
I really appreciate your help.
Gene
>>> storman AT US.IBM DOT COM 04/17/03 12:49AM >>>
The QUERY EVENT status is what you want to pay attention to.
Currently the SUCCESSFUL field in the SUMMARY table isn't useful for
determining the success of a client operation. This field corresponds
to
the "Normal server termination indicator" in the TSM server accounting
records, which basically says that the session between client and
server
ended normally.
Regards,
Andy
Andy Raibeck
IBM Software Group
Tivoli Storage Manager Client Development
Internal Notes e-mail: Andrew Raibeck/Tucson/IBM@IBMUS
Internet e-mail: storman AT us.eyebm DOT com (change eye to i to reply)
The only dumb question is the one that goes unasked.
The command line is your friend.
"Good enough" is the enemy of excellence.
Gene Walters <gwalters AT GWMAIL.STATE.WV DOT US>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
04/16/2003 07:50
Please respond to "ADSM: Dist Stor Manager"
To: ADSM-L AT VM.MARIST DOT EDU
cc:
Subject: Re: Query Event VS Successful Flag
Well, that is the only even I can see for that server. I ran another
query for the last few days on that server and here are the results:
Policy Domain Name: STANDARD
Schedule Name: SQLBACKUP
Node Name: WVWFISQLBACKUP
Scheduled Start: 04/14/2003 01:00:00
Actual Start: 04/14/2003 01:09:48
Completed: 04/14/2003 01:29:51
Status: Failed
Result: 402
Policy Domain Name: STANDARD
Schedule Name: SQLBACKUP
Node Name: WVWFISQLBACKUP
Scheduled Start: 04/15/2003 01:00:00
Actual Start: 04/15/2003 01:14:04
Completed: 04/15/2003 01:31:03
Status: Failed
Result: 402
Policy Domain Name: STANDARD
Schedule Name: SQLBACKUP
Node Name: WVWFISQLBACKUP
Scheduled Start: 04/16/2003 01:00:00
Actual Start: 04/16/2003 01:01:06
Completed: 04/16/2003 01:10:25
Status: Failed
Result: 402
and here is the only record that shows in the summary table on 4/16
for
that server:
START_TIME: 2003-04-16 01:03:56.000000
END_TIME: 2003-04-16 01:10:25.000000
ACTIVITY: BACKUP
NUMBER: 8715
ENTITY: WVWFISQLBACKUP
COMMMETH: BPX-Tcp/
ADDRESS: 129.71.217.20:4703
SCHEDULE_NAME:
EXAMINED: 8
AFFECTED: 8
FAILED: 0
BYTES: 460421272
IDLE: 8
MEDIAW: 0
PROCESSES: 1
SUCCESSFUL: YES
VOLUME_NAME:
DRIVE_NAME:
LIBRARY_NAME:
LAST_USE:
>>> storman AT US.IBM DOT COM 04/16/03 10:20AM >>>
The starting time from QUERY EVENT shows 01:01:06, while the starting
time
from your SELECT shows 01:03:56. Are you sure these two outputs are
for
the exact same event?
Regards,
Andy
Andy Raibeck
IBM Software Group
Tivoli Storage Manager Client Development
Internal Notes e-mail: Andrew Raibeck/Tucson/IBM@IBMUS
Internet e-mail: storman AT us.eyebm DOT com (change eye to i to reply)
The only dumb question is the one that goes unasked.
The command line is your friend.
"Good enough" is the enemy of excellence.
Gene Walters <gwalters AT GWMAIL.STATE.WV DOT US>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
04/16/2003 06:39
Please respond to "ADSM: Dist Stor Manager"
To: ADSM-L AT VM.MARIST DOT EDU
cc:
Subject: Query Event VS Successful Flag
I Run a variety of queries each night, one being the query event,
another being a select statement on the summary table. Here is a
sample
of what I get:
The query event:
Scheduled Start Actual Start Schedule NameNode Name
Status
---------------------------------------------------------------------------
04/16/2003 01:00:00 04/16/2003 01:01:06 SQLBACKUP
WVWFISQLBACK-Failed
The select:
ACTIVITY Server MB Duration(Min)
K/Sec SUCCESSFUL Start Time End Time
BACKUP WVWFISQLBACKUP 439.09 6 1155.86
YES 01:03:56 01:10:25
My question is: which one of these do I believe, the query event or
the info from the summary table. One says it failed, one says it was
successful?
Thanks
Gene
|
