New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Or probably is a machine with a kernel capable of both virtualization techs?
What host is this, @Zero?
@Aldryic TechieVPS
They might be running OpenVZ on one of their Xen nodes
Tokyo VPS with 600Mb/1.2GB Burst, 10GB HDD with 1TB of transfer for $5.95/mo ($60/Y) ?
@LES Yes, but in The Netherlands.
@dmmcintyre3 OpenVZ does not need it's own kernel to work?
http://wiki.openvz.org/Running_Xen_and_OpenVZ_in_the_same_host
Just do a "ps aux", if a bunch of strange processes in square brackets show up its probably xen, otherwise OpenVZ.
@dmmcintyre3 Ouch, I didn't know that.
@gsrdgrdghd Nothing strange there.
Ok, seems that they are running OpenVZ & Xen on the same machine and since the I/O is 19.1 MB/s I think is being oversold.
That's one way to oversell Xen ...
check if /proc/xen exists
Francisco
Yes, there is a directory called xen that contains these files:
balloon capabilities privcmd xenbus
paste balloon please
Francisco
Yes, that's OpenVZ running under Xen. I do this all the time on development containers so I don't have to buy a dedicated server.
@Francisco
Looks like it most likely is.
The whole node only has 6.7G RAM.
It's possible they're using XEN to give them a ghetto style remote KVM?
Francisco
It's not likely, it is. That's the OpenVZ kernel for Xen...
What you have here is OpenVZ in the cloud. You have a truly dynamically, scalable solution where node memory and disk space can be added ad-hoc (without downtime) across 5 locations, (8 more coming in the next few weeks). Server backups,security and redundancy are all abstracted. All storage is backed onto enterprise class SANS with significantly more redundancy than using only local storage solutions
If the hardware you VM is running on fails then it auto-migrates to another machine and restarts
i lol'd
@liam no, but you can still admire the water vapour that is left
@liam but that's exactly what 'cloud' is ;-)
-
Only an idiot would ever promise "100% uptime"
Cloud doesn't have a definition. The ones people accept are normally related to billing, in that people can up/down grade 'on the fly', normally just requiring a quick reboot and a partition resizing if it's storage.
My #1 concern with such a setup is always SAN performance and SAN's going R/O when they fail over. I can't imagine endangering customers data to such harsh rebuilds. I'm sure you guys/gals have been trying to improve that, but that's just how DRBD/etc handles rebuilds I guess.
Lordy, if VPS.net is claiming that's the meaning of cloud.... I can't finish this since there's no way to not sound like a complete arsehole.
Francisco
To me, cloud... is the fluffy thing that floats in the sky. It's not always up though, sometimes you get a blue sky with no clouds.
Understand i'm not taking a shot at their staff, skill or anything like that. What I'm taking a pot shot at is just SAN's in general. VPS.NET has staff/developers that will put together a system in an afternoon that'd take me days/weeks to come up with - If I even could.
Anyone that follows cloudfail, though, would see that the majority of the 'cloud' providers out there, especially VPS.NET have serious SAN issues. Yes, they're working to fix it but that was the same story last year and the time before that when they were trying to use ZFS I think.
SAN's are just trouble starters and a risky thing to get into. Gigecloud has had issues with their SAN's. Amazon has melted down for multiple days because of their centralized SAN (local storage was fine, of course).
Rackspace was smart in this all, they are like linode with a 'cloud like' environment, where it's just local storage with cloud provisioning ontop + tons of head room on each node.
Unless you have extremely deep pockets to get an enterprise SAN like steadfast I guess, you're going to have one hell of a time competing with your linux NFS/iSCSI setup. Sure, it might bench well when it works....when it works.
Francisco
I love the way rus foster always finds new ways to oversell like fk. Its truly amazing Credit to you Rus!
SAN aren't easy. SAN can be a nightmare :-)
But they are dependable, handle very high loads and you can clusterize HA services.
We have 2 kind of SAN which sometime got us some headache, but at the same time allow us to have record uptime for most of our services:
I must admit that the target of our services wasn't the low end market, else I hardly think we should have bought the fiber channel gears for cheap services