Comments
5 comments
-
Hi William,
This is something we discussed today and might change in a future build. Currently, as you've seen, we only show a warning if the copy fails altogether.
As an actual end user, would you like to see the yellow warning icon for transfers that completed successfully, but had handled network outage/jitters at some point?
Many thanks,
Robin -
It's fantastic to see that someone has been able to test this feature and that it works as advertised - there are too many users of our test system for me to be able to unplug it.
I'm going to find this feature invaluable in the operational environment as our txlogs backup directly to an offsite location.
Personally, I'd still like to see a warning that the job completed with network glitches. Mostly because it would alow me to spot potential network or NIC degradations before they become too much of an issue.
Andy -
RBA wrote:...
As an actual end user, would you like to see the yellow warning icon for transfers that completed successfully, but had handled network outage/jitters at some point?
Many thanks,
Robin
Hi Robin,
Yes I would like to see a warning, not an error. The yellow icon always tells me that things have finished, but something had a hiccup. I can then investigate the problems from there.
Regards
William -
Thanks for your feedback!
-
I have done some preliminary tests on network failures and have to agree, the methods employed work very well.
I would like to see that network issues were encountered during the process recorded somewhere in the GUI or the log.
Have mentioned it elsewhere but seems appropriate here too; I would like to see the progress made in the backup job logged. If the job does many databases and some got completed before the network error then identifying that they are complete, one has failed/suffered issues and others are (potentially) not backed up as the job exited after x retries...
Jonathan
Add comment
Please sign in to leave a comment.
I was in touch with one of your testers regarding network resilience and have now been able to test within a clustering environment.
Following setup:
Source Server
Windows 2008 + SQL Server 2008 (64 Bit single instance)
Target Server
Windows 2008 2 node active/passive failover cluster
I ran the backup of a 15 GB Database with compression level 4 (1.5GB - beats level 3 by 450MB!)
The copy over the network was "broken" via a manual cluster failover. After 30 seconds the copy job continued from where it stopped! :shock:
It did NOT start from 0KB, but had conveniently left a .PARTIAL file on the target disk. I did not expect that, maybe I should RTFM :oops:. This should definitely be a major marketing point IMO.
Now comes the next nice bit. I refailed the cluster over and the transfer started off from previous mid-transfer again. Multiple fails in a row really do not upset the copy process.
What didn't make sense, was that the history log for the backup showed no signs of network problems. The job finished as successful and the log created just said that the copy job was a success. Have I missed something here, or is that desired behaviour?
Regards
William