| OpenVz HA with common repository without DRBD [message #51134] |
Thu, 20 February 2014 14:09  |
topgunx1
Messages: 2
Registered: February 2014
|
Junior Member |
|
|
Hello everybody,
First of all thanks for this forum and your precious help that you would give me!
I have a question for you.
This is my scenario:
2 Nodes (Host A and Host B) on which there is installed OpenVz,that I will link via Fiber Channel to a SAN (common repository).
If one of 2 nodes crashes, I would want that the VE of the node down can migrate towards the second host.
Another requirement is that I don't want to use DRBD because I want to save all on my SAN.
How can realize this in practise?
Is OpenVZ thought for do this?
Does it require the use of GFS?
And the last question: is there anyone who have realized this kind of structure?
Thanks in advance.
|
|
|
|
| Re: OpenVz HA with common repository without DRBD [message #51153 is a reply to message #51134] |
Wed, 26 February 2014 19:01  |
Paparaciz
Messages: 302
Registered: August 2009
|
Senior Member |
|
|
if anyone did this it was completly custom build.
but I don't think that it would be hard to do so.
also it depends what you mean HN is crashed.
some monitoring scripts can just check other host and in case of some failure mount needed partition/partitions, start CT's
in case HN is predicted failure, than script which checks self HN can than checkpoint, copy/move checkpoint to other HN, stop vz, unmount effective partitions, init needed actions in other HN to work properly(mount partitions?), and than do restore of CT in other HN. as for checkpointing/restoring in other HN take a look to "vzmigrate" script.
of course you need somehow sync CT config files.
btw, i'm not in openvz team, but saw they are advertising as a demo of parallels cloud storage solution:
https://wiki.openvz.org/Parallels_Cloud_Storage
but for me it has very unpleasant limitations, as example only 100GB for CT's
|
|
|
|
|
|
OpenVZ Forum
Members
Search
Help
Register
Login
Home
