Unsolved
2 Posts
0
783
June 4th, 2020 11:00
how and where to calculate average backup speed of backup and restore in Networker Backup server?
how and where to calculate average backup speed of backup and restore in Networker Backup server?
No Events found!



bingo.1
2.4K Posts
0
June 4th, 2020 13:00
What about using the same method which you use for any other data transfer: divide the amount of data by the duration of the backup?
A general calculation/prediction is almost impossible - only for a single process you could run a backup and clearly determine the bottleneck. However, NW is considered to do multiple things at the same time and of course the scenario changes, especially if you have larger environments. And co course, we do not know your environment.
So I can only encourage you to test, test, test ...
Bala Msd
2 Posts
0
June 4th, 2020 23:00
Thanks for the reply, we have huge set up around 26K plus clients and 20+ backup servers .. can be calculate each by doing nsrwatch and we have script to see the backup flow, but my requests are to calculate average backup speed for last at least a week or month and restore speed. Is there any common way to do ?
bingo.1
2.4K Posts
0
June 5th, 2020 02:00
For backups you can get very precise information from "mminfo ... -r ..." . "sumsize(#)" will show you the size of a save set precisely to the byte. "sscomp(#)" & "sscreate(#)" will return start and stop time. Along with other parameters, you can create a very precise report for the save sets which have been created during the requested period of time. Export the result into a csv table.
Then, per script, convert the time stamps to Epoch format and calculate the difference - this will give you the duration. Now you have everything ready to calculate the speed. If you now sort the table by whatever column you want, you can easily find and sort the results or prepare a statistic.
Now - NMC reports or data protection advisor may also give you good results. I do not use them as I do not know how they do their calculation. But I can easily verify/adopt my own method .
Unfortunately, I have never looked into recover processes so I cannot propose anything right away.