Xtravirt Virtual SAN Appliance

[admin]

Hi Guys,

Over at Xtrairt we've just released a new free virtual appliance that I thought people might find interesting. It lets you use the local storage on two ESX hosts to create a virtual SAN. I've attached an overview diagram to explain how it works.

http://www.xtravirt.com/index.php?option=com_remository&Itemid=75&func=fileinfo&id=29

Summary: The Xtravirt Virtual SAN (XVS) appliance for VMware ESX3 Server is a free solution to provide the benefits of shared VMFS storage without the cost of a SAN – this allows the utilisation of otherwise unused local storage in the ESX server to facilitate enterprise level features such as vMotion, DRS and HA normally only available through the use of a shared storage device. All volume data is synchronously replicated between hosts, providing full fail-over capability with data integrity in the event of host, disk or appliance failure. The appliance is menu driven and has been designed to be as easy to configure as possible, and full documentation on the implemenation process is provided.

Questions/feedback welcome in this thread.

Cheers,

Alex

[admin]

We do offer commercial support, please see here for more details: http://www.xtravirt.com/index.php?option=com_content&task=view&id=95&Itemid=123 (or email sales@xtravirt.com).

Thanks,

Alex

[rocker77]

Did anybody test complete recovery of vSAN? I try import new XVS appliance, add fresh disk, but can´t sync data from working node to new installed??

Thanks

[admin]

I've tested this and it worked fine for me, if you send me an email with your config/some screenshots I'll try and figure out what's going wrong.

Regards,

Alex

[rocker77]

I send you email about support. I will test it once more. Thank you.

[TomHowarth]

Thread moved to a more appropiate forum

Tom Howarth

VMware Communities User Moderator

[JRink]

I'm trying to understanding how this works... Couple questions.

If I have my ESX host with a 72GB RAID 1 volume (hosting ESX OS and a small Datastore1) and a 280GB RAID 5 volume, I'm assuming I would want to setup the SAN VM on my small Datastore1. I'd then create an additional hard disk for the SAN VM of, for example, 250GB. I'd want this 250GB hard disk to be located on my RAID5 volume. But, I'm not sure I understand this correctly when I read your MS Word document (it mentions using the "Store with virtual machine" option when adding the 2nd disk). I think I'm a bit confused as I never setup a VM with indepedent disks either and maybe that's where I'm confused. Trying to grasp this...

Also, about the replication feature between local disks on each ESX host... If my production VM are stored in the LUN off the SAN VM, and the SAN VM is continuously replicating the data from one ESX DAS to the other ESX DAS, how does this effect performance? Wouldn't this replication also hinder the performance of my VMs? (this isn't an Extravirt question really, just a general question about this kind of technology).

JR

[stormlight]

Any answer to the above question from anyone who has tested this?

"Also, about the replication feature between local disks on each ESX host... If my production VM are stored in the LUN off the SAN VM, and the SAN VM is continuously replicating the data from one ESX DAS to the other ESX DAS, how does this effect performance? Wouldn't this replication also hinder the performance of my VMs? (this isn't an Extravirt question really, just a general question about this kind of technology)."

[JRink]

I can't seem to figure this out. I'm trying to test this but am running into a problem when I try to get my VM a heartbeat IP address. I am trying to use 10.0.0.5 for the IP address but it says, "Invalid Option" when I do this.

I'm working on test boxes, so they aren't configured exactly as I like but... my configure is with a single NIC in each ESX host.

My 1st ESX server has a single vSwitch0 which has ServiceConsole1 at 192.168.1.5/24, ServiceConsole2 at 10.0.0.11/8, VMKernel at 10.0.0.1/8.

My 2nd ESX server has a single vSwitch0 which has ServiceConsole1 ata 192.168.1.6/24, ServiceConsole2 at 10.0.0.12/8, VMKernel at 10.0.0.2/8.

Why is it erroring when I try to set a custom IP address? If I manually enter 192.168.1.x/24 range for the IP, then it accepts it, but I really don't want to use 192.168.1.x/24 for the network. I typically set all my backend iSCSI networks up on 10.0.0.0/8.

Any advice here? Frustrated.

JR

[JRink]

I just figured out the problem. The VM does NOT allow you to enter IP addresses that have "0" in an octet.

I have seen this problem before on a UPS unit's SNMP management card. You couldn't configure the IP if it had a "0" octet. Likely a programming issue.

Anyways, I changed my 10.0.0.x network to a 10.1.1.x network and it finally took the values.

JR

[JRink]

More issues with this....

It seems like when rebooting either of my ESX hosts, whether sometimes doing clean shutdowns or simulating a server crash by pulling the plug, the 2 LVS nodes do NOT want to communicate anyone and do not even offer up iSCSI. When I try started the LVS nodes, the iSCSI storage never comes online. Furthermore, going into the LVS node menu and selecting Start Node never ends up synchronizing with the other node. The only thing it says is

" 0: cs:standalone st: secondary/unknown ds:consistent/dunkown r----

ns:0 nr:0 dw:0 al:0 bm:0 lo:0 pe:0 ua:0 ap:0

resync: use: 0/31 hits:0 misses:0 starving:0 dirty:0 changed:0

act_log: used:0/127 hits:0 misses:0 starving:0 dirty:0 changed:0

BOTH nodes are showingthe identical thing. Again, the iSCSI storage is not online.... The only way I can seem to fix this is to reset both nodes and reconfigure both nodes (I can't find a way OUT of that screen when it's trying to sync... ctrl-c, etc. nothing worked to get back to the menu...).

ON a side note, aside from THIS issue, I am having problems wiht the sole VM that is running on Node1. It's a Win2003 VM and is constantly crashing and becoming unpingable for no apparant reason.....

Is there ANY support available for this? I know it's free.. but I would like to introduce this to a couple clients, however I can't really do so if I'm having problems even getting this up and running.

Please advise. a bit frustrated with this...

JR

[admin]

Hi JR,

I just figured out the problem. The VM does NOT allow you to enter IP addresses that have "0" in an octet.

I have seen this problem before on a UPS unit's SNMP management card. You couldn't configure the IP if it had a "0" octet. Likely a programming issue.

Anyways, I changed my 10.0.0.x network to a 10.1.1.x network and it finally took the values.

JR

I can confirm this is a bug. I'll get it fixed shortly and put out a new release.

More issues with this....

It seems like when rebooting either of my ESX hosts, whether sometimes doing clean shutdowns or simulating a server crash by pulling the plug, the 2 LVS nodes do NOT want to communicate anyone and do not even offer up iSCSI. When I try started the LVS nodes, the iSCSI storage never comes online. Furthermore, going into the LVS node menu and selecting Start Node never ends up synchronizing with the other node. The only thing it says is

" 0: cs:standalone st: secondary/unknown ds:consistent/dunkown r----

ns:0 nr:0 dw:0 al:0 bm:0 lo:0 pe:0 ua:0 ap:0

resync: use: 0/31 hits:0 misses:0 starving:0 dirty:0 changed:0

act_log: used:0/127 hits:0 misses:0 starving:0 dirty:0 changed:0

BOTH nodes are showingthe identical thing. Again, the iSCSI storage is not online.... The only way I can seem to fix this is to reset both nodes and reconfigure both nodes (I can't find a way OUT of that screen when it's trying to sync... ctrl-c, etc. nothing worked to get back to the menu...).

ON a side note, aside from THIS issue, I am having problems wiht the sole VM that is running on Node1. It's a Win2003 VM and is constantly crashing and becoming unpingable for no apparant reason.....

Is there ANY support available for this? I know it's free.. but I would like to introduce this to a couple clients, however I can't really do so if I'm having problems even getting this up and running.

Please advise. a bit frustrated with this...

Sorry for the delay spotting your posts, and some issues getting the XVS to work. Standalone mode is a state where the nodes are refusing to speak to eachother for some reason, usually a split-brain scenario. This seems likely because both nodes believe they are in a consistent state so you end up with a dead-lock. The reason the iSCSI storage won't come up is because in secondary mode the storage is read-only. You'll need to run some commands at the console to fix this state.

I've not seen these issues in my dev setup and have been using the XVS to host lots of test Windows VMs for some time without issue, so I'm sure we can fix the problems. Can you send me some details of your setup, the hardware you're running, exactly how you failed the node, etc, to support@xtravirt.com? I'll do my best to try and get you up and running.

Cheers,

Alex

[JRink]

Sent a couple emails tonight per your request...

JR

[SimonG-ioko]

I also get the same problem, if I reboot the 2nd ESX box (or even just the SAN VM) it won't come back up (pic attached)

Let me know if/what further info you need?

Cheers

[pauldavidbarker]

I also get this problem has then been an update I missed

[JRink]

Not that I'm aware of. I sent their support a couple of emails about 10 days ago with lots of information, but haven't heard back at all.

JR

[admin]

Sorry for the delay replying guys, been on holiday in sunny Wales (har har).

JRink - replied to all your emails, looks like a network connection issue causing the nodes to go offline to prevent split-brain.

SimonG - the node being in standalone happens for one of two reasons, the network connection cannot be established or both nodes think they hold consistent copies of th storage so refuse to connect (again to prevent split brain). How did you shut down the appliance on the second ESX server? Did you shut down the services as per the documentation, or just pull the plug?

If you could email the contents of /var/log/mesages from the XVS nodes to support@xtravirt.com I'll take a look and try and work out what's going on.

Cheers,

Alex

[JRink]

Alex,

I'm not sure how it could be a network connection issue when the 2 hosts are connected via crossover cable. ??

I will take some time tonight to go over your email replies more thoroughly and get back to you with more information.

JR

[jgal]

Jrink you mind posting some of those emails. We are having a problem using cross over cables as well. After the initial setup it takes a few reboots or to wait some time until we can start the first sync

If we try to sync after configure it never connects. I do rember reading something in the logs that says it didnt want to be the primary. We have full connectivity and the only thing that fails is the ntp sync since its on a crossover network between 2 host.

[rvsharpe]

Is there any update regarding using Xtravirt with IP addresses with a "0" octet?

rvs

[rvsharpe]

Never mind...I just changed the network to a 192.168.1.x network. Installation was fine after that. As stated previously in this thread, the application seems to have problems when resyncing after rebooting or shutting down and re-starting one of the nodes..neither node will communicate with each other and a reconfiguration is needed, at least in my case..