StarWind Software

Beta-2 is scheduled to be out this week. Yes.

Omnividente wrote:Wnen will the new beta realise? Loopback acceleration on windows Server 2012 r2 will be fix?

Omnividente wrote:Wnen will the new beta realise? Loopback acceleration on windows Server 2012 r2 will be fix?

May I ask you about quick test?
Crate RAM disk using any emulation software (for example http://www.starwindsoftware.com/high-pe ... k-emulator - it is free product, or http://memory.dataram.com/products-and- ... re/ramdisk )
1) Format the disk and launch ATTO Disk benchmark against it.

On the same machine install StarWind with "Loopback connection accelerator" enabled.


Create StarWind ImageFile device located on that RAM disk device.
2) On the same machine launch MS iSCSI initiator, add 127.0.0.1 as a target portal, connect the discovered ImageFile. Format the disk and launch ATTO Disk benchmark against it.

Publish here screenshots with the resuls for 1) and 2)

Thank you!

...also we would need StarWind logs from both of the cases. Thanks!

Hello,

I am working with your beta right now, namely the replication features. I'm not sure if the functionality was removed or just moved somewhere else, but I can't seem to find a way to specify the target name of my replica (which was available in the last version).

I've tried disconnecting the image file from the created replica target and attaching to a new one, but that doesn't work. Furthermore, if the image file is already attached to a target, it doesn't seem to be able to replicate either.

If this feature has been removed entirely, PLEASE consider re-adding it as I do not like the way it just automatically names my replica target something ugly.

Furthermore, I do not like the fact that I can't choose the destination of my replica .img file either... It just drops it in the root volume, which makes keeping track of these things an absolute pain.

Please also bring this feature back, along with specifying a target replica name... They are the absolute most important features with this replication solution for me.

May I ask you about quick test?

Windows Server 2012 r2
1.RamDIsk 2.Image on RamDisk



Reslult better than 3 days ago http://www.starwindsoftware.com/forums/ ... tml#p19597 I do not understand why

I tested loopback on my HDD

1. WIndows Server 2012 R2

1.HDD 2.Virtual disk-thick provisioned on this HDD and 127.0.0.1 target portal without cache


2. WIndows Server 2012 (The old server is more slow disks)

1.HDD 2.Virtual disk-thick provisioned on this HDD and 127.0.0.1 target portal without cache



I see that the loopback acceleration works best on Windows 2012 than Windows 2012 r2.

questions:
1.How can I add a cache on HA partner? Now I can only do cache when the image added, after the added of a partner? cache is only on the first node.
2.How can i mark node "non optimization for alua"?

Testing current v8 beta...
Configured an standard .iMG with 3 GB WB cache.

If you disconnect LUNs and try stopping the service, it take long and forver (refuse to stop) and you must kill task manually with Windows Task Manager.

Hope fixing in next release...

Mardes,
Thank you for the information.
There was a bug with flushing cache on device close. It is fixed and will be released with the next build of beta version.

Ran into that one myself. after tying to delete a device with 20GB+ cache. Now using 3 lsfs devices with 8GB cache each, very good performance. Looking forward to testing the beta2. I want to use all the free mem on the system

I guess this is too late for feature requests, but one feature that will be really nice will be to group initiator access. Basically allow multiple initiators to be marked as single host and then allow multiple hosts to be grouped and then allow these groups to assigned access rights. this way I can group all my Hyper-V hosts in one group and vSphere hosts in one group. Just makes life a bit easier.

Omnividente wrote:

May I ask you about quick test?

I tested loopback on my HDD
1. WIndows Server 2012 R2
1.HDD 2.Virtual disk-thick provisioned on this HDD and 127.0.0.1 target portal without cache

What is the disk model? Is it SSD? Is it single HDD or several HDDs in RAID?

Bohdan (staff) wrote: What is the disk model? Is it SSD? Is it single HDD or several HDDs in RAID?

RAID 0

Omnividente wrote:

Bohdan (staff) wrote: RAID 0

1) SATA or SAS?
2) Hardware or software RAID?
3) What is the RAID0 stripe size?

Bohdan (staff) wrote:

Omnividente wrote:

Bohdan (staff) wrote: RAID 0

1) SATA or SAS?
2) Hardware or software RAID?
3) What is the RAID0 stripe size?

1.Sata
2.SoftWare
3.64kb

jwordtmann wrote:Hello,

I am working with your beta right now, namely the replication features. I'm not sure if the functionality was removed or just moved somewhere else, but I can't seem to find a way to specify the target name of my replica (which was available in the last version).

I've tried disconnecting the image file from the created replica target and attaching to a new one, but that doesn't work. Furthermore, if the image file is already attached to a target, it doesn't seem to be able to replicate either.

If this feature has been removed entirely, PLEASE consider re-adding it as I do not like the way it just automatically names my replica target something ugly.

It is fixed. The beta build will be updated by the end of the week. I'll post the notification in this topic.

1.How can I add a cache on HA partner? Now I can only do cache when the image added, after the added of a partner? cache is only on the first node.
2.How can i mark node "non optimization for alua"?

?)