Unsolved
This post is more than 5 years old
5 Posts
0
6443
failed and aborted save set - mminfo
Hi,
How can we pull the only save set which were failed (some part backup) or aborted through mminfo command.
Vineel
Unsolved
This post is more than 5 years old
5 Posts
0
6443
Hi,
How can we pull the only save set which were failed (some part backup) or aborted through mminfo command.
Vineel
Top
ble1
2 Intern
2 Intern
•
14.3K Posts
0
March 31st, 2016 04:00
MDB contains data about media database. This means that for disk backups you won't have this data as such data is removed upon failure (to preserve disk space and not trash it with useless data). So, with mminfo you can see failures only for data which is left on media after failure/error (eg. on tape). Such save sets are marked as aborted. Check mminfo manual for ssflags and sumflags fields.
StuartWhitby
45 Posts
0
March 31st, 2016 06:00
However, if you want to pull information from the media db about aborted savesets, you can't query this directly (surprisingly - you can match "recoverable" or "incomplete", but not "aborted" using mminfo). As such, to see all savesets which are aborted in the media db, it'd be:
mminfo -av -r ssid,sumflags | grep a
(or "findstr a" if on Windows).
That gives you the list of SSIDs. Depending on what you want to use this info for, you can query individually on those or improve the above search query ("mminfo -avot -r client,name,sscreate,sumfags | grep ca$" maybe).
Or you can get DPA if you want to get fancy with reporting.
Shubhadip_Datta
26 Posts
0
April 5th, 2016 04:00
Hi Vineel,
Since mminfo never capture the data for failed/aborted save set record, you can't report that in mminfo query. But using ssflags filed in mminfo output you might get lot of import information/state of a save set.
Please refer below link for more details to get the details of ssflags -
https://community.emc.com/docs/DOC-24741
Now coming back to your actual requirement , yes you can report failed/aborted backup jobs using jobquery. Below I am providing some small snippet for that for a specific client -
jobquery> . type:save job;host:xyz
Current query set
jobquery> show completion status;host;job state;name;saveset name;start time
jobquery> pr
completion status: succeeded;
host: xyz;
job state: COMPLETED;
name: /var/log;
saveset name: /var/log;
start time: 1459843226;
completion status: abandoned;
host: xyz;
job state: COMPLETED;
name: /opt;
saveset name: /opt;
start time: 1459756806;
completion status: succeeded;
host: xyz;
job state: COMPLETED;
name: /usr/sap/HRB;
saveset name: /usr/sap/HRB;
start time: 1459849779;
completion status: succeeded;
host: xyz;
job state: COMPLETED;
name: /var;
saveset name: /var;
start time: 1459843226;
completion status: abandoned;
host: xyz;
job state: COMPLETED;
name: /home;
saveset name: /home;
start time: 1459670407;
Now from above illustration you can easily report the abandoned/failed save job as per your requirement. You can use your own scripting skill now to format the output or restrict the search limit.
Note:
i) Remember the job state indicates the over all save job running state - COMPLETED/ACTIVE/SESSION ACTIVE etc
Where completion status will tell you the exit status of that save job - succeeded/failed/abandoned etc.
ii) Start time will report the time in unix timestamp format , using your own script you can convert those in readable format.
Thanks & Regards,
Shubhadip Datta
bingo.1
2.4K Posts
0
April 5th, 2016 06:00
Use mminfo
- For disks look for 'suspect' save sets.
- For tapes, look for 'aborted' save sets.
That's it
bingo.1
2.4K Posts
0
April 5th, 2016 07:00
Correct. That's why I remove the suspect disk backups before I run into clone problems.
ble1
2 Intern
2 Intern
•
14.3K Posts
0
April 5th, 2016 07:00
suspect for disk makes sense only to trace those that has issue during read operations (eg. cloning or restore) - aborted once are not covered in such case as they are removed.