This post is more than 5 years old
71 Posts
0
2489
Difference Between mminfo -a and mminfo -av
Hello
I hope this isn't as stupid a question as it seems
If I run the query:
mminfo -a -r "totalsize" | awk '{sum=sum+$0};END{print sum/1024/1024/1024/1024}'
I get a figure of 219.841 Tb. However, if I run the query:
mminfo -av -r "totalsize" | awk '{sum=sum+$0};END{print sum/1024/1024/1024/1024}'
I get a figure of 1087.28 Tb.
Narrowing the time down to the last 6 hours and expanding the -r switch, it looks like the v (verbose) switch adds the index: and bootstrap: savesets - but that would mean that all my indexes and bootstraps are four times the size of my backups.
Can someone please confirm if this is the case? Also, when reporting the total size of all backups, which switches should be used - the -a or the -av switches?
Incidentally, we have approximately 750 LTO3 tapes. If we take the capacity of each tape to be 400 Gb, that equates to almost 300 Tb, therefore it looks as if the -a figure is more accurate.
Brian
bingo.1
2.4K Posts
0
March 7th, 2014 07:00
You better be careful when you use the preconfigured reports - not all details may be fully documented.
The best solution is to use "mminfo -q -r " to define your own reports.
'totalsize' is the correct parameter to get the amount of data. However, your calculation is not correct.
Your formula will show you the size in Tebibytes, not in Terabytes. For more details, please see the web (WIKI).
Lucky85
173 Posts
0
March 7th, 2014 05:00
Part of cmd guide:
Specifying the – v flag prints aborted, purged, incomplete and recoverable save sets in
addition to the complete, browsable save sets printed by default. The – v flag also
causes three additional fields to be displayed: the creation time, the internal save set
identifier (ssid), and two flags. One character is used per flag.
Regards
Lukas
Brian91_4544b4
71 Posts
0
March 10th, 2014 04:00
Thanks for your reply, bingo.
Playing around with the -q queryspec, the code:
mminfo -a -r "totalsize" -q "savetime >= '01 January 2007', savetime <= '01 April 2014'" | awk '{sum=sum+$0};END{print sum/1000/1000/1000/1000}'
reports 379.063 Tb, whereas the code:
mminfo -av -r "totalsize" -q "savetime >= '01 January 2007', savetime <= '01 April 2014'" | awk '{sum=sum+$0};END{print sum/1000/1000/1000/1000}'
also reports 379.063 Tb. In other words, you're correct in saying that you should use an absolute queryspec as without this, additional things seem to be getting added. (Our long-term save sets are set for 6 years, so I know that there are no save sets more than 6 years old. I'll create a more elegant solution for the date when I get time to change my script.)
As for reporting in Tebibytes and not Terabytes, what can I say!?! I simply based my code on stuff someone else had written and I didn't think to query it. And the really embarrassing thing is that I've been using this for years...
Brian
ble1
2 Intern
2 Intern
•
14.3K Posts
1
March 10th, 2014 08:00
What NW is using is /1024/1000/1000 (I agree with respect to that that above value should be 751G rounded)
bingo.1
2.4K Posts
0
March 10th, 2014 08:00
With respect to the calculation ... believe me - you are not alone. I got trapped myself a few years ago. And a lot of others are suffering as well.
BTW - NetWorker uses a 'mixed' calculation. For example:
C:\>mminfo -q "client=client,name=H:\,level=full,savetime>=today" -r totalsize,sumsize
total size
768784472460 750 GB
C:\>
The 'sumsize' value is obviously calculated as follows: sumsize = totalsize/1024/1000/1000
These mixed (mis) caclulations are also common but technically incorrect.