The environment is Progress 11.5 64 bit Windows server 2012
We are testing moving our development environment to the Azure cloud. We're seeing some performance issues with DB restore time, file transfers etc.
The original environment is a VM running the same OS and connected to a SAN. I am lacking in details about the SAN but suffice it to say that there was nothing special about its performance or configuration. Performance has always been acceptable. I believe it's probably RAID 5 with a lot of cache.
In the old environment, a restore of a 100gb database takes about 18 minutes. On Azure, about 55. This is with the space for the extents already allocated. We've been working with a group in our company that's taken responsibility for Azure, they've done a lot of changes, some have improved performance a bit but there's still a very large difference. Some other operations are not quite as proportionally slow but they're all slower.
At the moment, I believe we're on a DS14 server and connected to P30 storage. This seems like overkill for what we need but they keep trying different things and that's where we are right now.
Has anyone done this in Azure? Any suggestions for proper setup of the Azure VMs and storage?
Thanks,
Tom
i have no experience with azure.
most likely the slower restore speed is caused by slower storage, shared with who knows how many others.
we in the Progress-BravePoint MDBA group do various test to measure storage system speeds. an easy one is this:
procopy $DLC/demo foo
proutil foo -C truncate bi -bi 16384 -biblocksize 16384
time proutil foo -C bigrow 2
Repeat several times to get a consistent result.
On halfway decent storage, this will take about 4 seconds. We have neasured times as high as 2 minutes. Those customers are constantly complaining about performance. Anyhow, you can do that test to compare your original environment with the new one.
> On Nov 30, 2015, at 8:11 PM, tbergman wrote:
>
> Update from Progress Community [https://community.progress.com/]
>
> tbergman [https://community.progress.com/members/tbergman]
>
> The environment is Progress 11.5 64 bit Windows server 2012
>
> We are testing moving our development environment to the Azure cloud. We're seeing some performance issues with DB restore time, file transfers etc.
>
> The original environment is a VM running the same OS and connected to a SAN. I am lacking in details about the SAN but suffice it to say that there was nothing special about its performance or configuration. Performance has always been acceptable. I believe it's probably RAID 5 with a lot of cache.
>
> In the old environment, a restore of a 100gb database takes about 18 minutes. On Azure, about 55. This is with the space for the extents already allocated. We've been working with a group in our company that's taken responsibility for Azure, they've done a lot of changes, some have improved performance a bit but there's still a very large difference. Some other operations are not quite as proportionally slow but they're all slower.
>
> At the moment, I believe we're on a DS14 server and connected to P30 storage. This seems like overkill for what we need but they keep trying different things and that's where we are right now.
>
> Has anyone done this in Azure? Any suggestions for proper setup of the Azure VMs and storage?
>
> Thanks,
>
> Tom
>
> View online [https://community.progress.com/community_groups/openedge_rdbms/f/18/t/21692]
>
> You received this notification because you subscribed to the forum. To stop receiving updates from only this thread, go here [https://community.progress.com/community_groups/openedge_rdbms/f/18/t/21692/mute].
>
> Flag [https://community.progress.com/community_groups/openedge_rdbms/f/18/t/21692?AbuseContentId=638aea3d-2ee2-465c-8110-a73e035c9021&AbuseContentTypeId=46448885-d0e6-4133-bbfb-f0cd7b0fd6f7&AbuseFlag=true] this post as spam/abuse.
Of course the next thing that happens is that the vendor tries to blame the problem on Progress... which is why it is handy to have a "dd" command (or something similar) to offer them as a nice "neutral" test which will show the poor performance of their hardware/vm/san/nas thing.