I've read several comments stating that file i/o mode should be used rather than block i/o when using iscsi targets with vmware. Does this recommendation still hold true?
Printable View
I've read several comments stating that file i/o mode should be used rather than block i/o when using iscsi targets with vmware. Does this recommendation still hold true?
Yes this still holds true, If you create an ISCSI volume in block I/O, a pop up window will advise you to switch to file I/O if using the volume with VMware.
I have been using VMWare with my ISCSI targets in file I/O without any problems.
Thanks for the responses.
I don't remember seeing any popups like you describe.... would that popup happen when in the DSS web interface, or are you referring to the vmware vi client?Quote:
Originally Posted by toepytoe
The pop up wil appear if you are in block I/O when creating a new ISCSI volume in the DSS web interface
I guess I must've switched to file i/o a few weeks ago when I first started testing so I forgot about that pop-up. I just changed back to block i/o, created a new iscsi volume, and saw that pop-up.
Thanks.
Sorry for the thread necromancy, but...
When presenting an iSCSI LUN to a VMware client using the MS iscsi software in server 2008... should I be using block I/O ?
Yes block IO
Can someone please explain why?
And what are the benefits of using file mode instead of block mode?
And is this recommendation for iSCSI volumes that will be used for VMFS volumes that will contain VM's?
Or is the recommendation for iSCSI volumes what will be used within VM's via a iSCSI initiator?
Link below provides information on Block IO and File IO, we recommend to use Block IO for VMware as this is what we certified the DSS V6 for the iSCSI volumes.
http://www.kb.open-e.com/File-IO-Or-Block-IO_342.html
So the prompt when creating a volume that says you guys highly recommend File I/O for VMWare is not correct based on current findings?
In the DSS V6 latest version this can go both ways, its just that we certified with the Block IO.
I am using open-e on a server that I am attempting to turn into a SAN. My vm host servers are running 2008 Server R2 with Hyper V (not the core installation I installed the OS).
I just created Volume Groups. I want to create disk partitions and put a virtual machine on each one. I want to create iscsi connections so that the VM is on the open-e storage server and my other server with hyper V can connect to it with iscsi connection.
Should I use File or Block I/O for a Hyper V setup? and if it is File what speed it shows slow, medium fast?
Use Block IO instead of the File IO as we have seen better performance.
What does one do if they have already set up esx servers with file i/o before Open-e V6 was certified with ESX4, before this file i/o was the best way to go.
Is there a way to change this to block i/o
There is no way you can change from File IO to Block IO, the only way is to create a new volume and copy over the data.
But using File I/O with plenty of RAM and fast enough CPU (Xeon QC) should benefit from the filesystem caching to improve read speed? So I don't understand how Block I/O would be faster than File I/O in a virtualized environment like ESX or XenServer? We are currently building two systems (one failover) with 5 SAS 15k and 5 SATA disks, of which both arrays will be an iSCSI target for XenServer hosts. Until now, we were convinced File I/O is the way to go. Could you explain a little bit more what new insights you have on this, since before File I/O was recommended?
Thank you.
This was because in the past with DSS V5 we had IET for the Target solution and now we use for the default SCST and in the next release the new version of SCST will be out.
http://www.kb.open-e.com/File-IO-Or-Block-IO_342.html
Use the Block IO as we certified with VMware EXS 4.0 with Block IO and the message will be removed from the GUI in the next release, File IO as stated on the link is adding a file
system layer and best to use Block IO as we certified it with VMware.
use block i/o for iscsi if you want to be able to use the volume replication & automatic failover features of open-e, file i/o initialisation is not necesary and the difference in performance is very small