iSCSI Failover

Goodmorning,

With our setup here, we are trying to build the following configuration:

Primary node -> Secondary node

Master target 0 -> Replicated Master target 0 (slave target)
Master target 1 -> Replicated Master target 1 (slave target)
Replicated Master target 2 (slave target) -> Master target 2
Replicated Master target 3 (slave target) -> Master target 3


So we have 2 targets on the primary node, these targets will be replicated on the secondary target. In case the primary node fails, the data on the targets will become available from the secondary target.

On the other hand, if the secondary node fails, we want to have the data on the targets available from the primary node.

Is this possible? Or can iSCSI Failover only be used for primary to secondary and not back?

Greetings!

hi,

when your primary is back and running from the secondary start sync volumes and after it finishes press the failback button. there is only manual failback possible.

melik

Quote:

---

Originally Posted by melik

hi,

when your primary is back and running from the secondary start sync volumes and after it finishes press the failback button. there is only manual failback possible.

melik

---

Melik,

This is not exactly what i ment.

We have got 2 servers set up here, both running open-e with 1 iSCSI target, replication enabled. iSCSI Failover is also set-up, it works great, even with a virtual machine running at the moment the primary node fails.

But we wanted to have a 2 way iSCSI Failover, with on both servers 1 target and 1 replication target. At the moment node 1 fails, node 2 will take over the target from node 2. On the other side if node 2 fails, node 1 will take over the target from node 2.

Is it possible to set this up, or is it only possible to have 1 server set up as a primary and 1 as a secondary node?

Greetings!

Melik,

This is not exactly what i ment. I think my question was a bit off, so i will ask it again :) my bad!

We have got 2 servers set up here, both running open-e with 1 iSCSI target, replication enabled. iSCSI Failover is also set-up, it works great, even with a virtual machine running at the moment the primary node fails.

But we wanted to have a 2 way iSCSI Failover, with on both servers 1 target and 1 replication target. At the moment node 1 fails, node 2 will take over the target from node 2. On the other side if node 2 fails, node 1 will take over the target from node 2.

Is it possible to set this up, or is it only possible to have 1 server set up as a primary and 1 as a secondary node?

Greetings!

Currently it is not possibly.

I have made an image which hopefully will explain it a bit better.

http://img651.imageshack.us/img651/6...openesetup.jpg

This is how our customer wants to have it!

You are describing an active/active configuration. Unfortunately, this won't work with automatic HA, but it may work if you can deal with manual failover.

enealDC is correct you would want to use only volume replication side of it.

Quote:

---

Originally Posted by To-M

enealDC is correct you would want to use only volume replication side of it.

---

So it is not possible to use a Failover configuration for active/active?

Because the whole idea was to set up a configuration in which the 2nd server will take over the tasks of the first server and vice versa. This is not possible through volume replication because the server will than only backup the data and client computers wont be able to access the storage when there primary server is down.


Greetings

Quote:

---

Originally Posted by Slijfer

So it is not possible to use a Failover configuration for active/active?

Because the whole idea was to set up a configuration in which the 2nd server will take over the tasks of the first server and vice versa. This is not possible through volume replication because the server will than only backup the data and client computers wont be able to access the storage when there primary server is down.


Greetings

---

Active / Active is not possible yet......

In failover, if Pri goes down, Sec WILL take over, no problem.
When Primary comes back you have to fail back... manually.
The only issue here is the manual part of failing back...... but keep in mind, the hosts will continue to use the targets the whole time, with no problem.

If you are willing to make an investment in some scripts and learning the API, you can have your cake and eat it too :)

1. Externalize the HA (e.g. the virtual ip) using a load balancer
2. Create a script or program that leverages the API to cause the volume replication to "flip"

Open-E, the API should allow for one to control volume replication and iSCSI Targets correct?

Quote:

---

Open-E, the API should allow for one to control volume replication and iSCSI Targets correct?

---

Yes, and No.....
Target manipulation can be achieved in the API, but the replication switching can be a bit tricky.. I havent tested it, but it may be possible.
got a test box around, and want to test this, i can throw ya some commands if u want.....

I'm going to take you up on that offer next week. I need to do some prep of my dev environment.

I have some more but let me know if this helps.

create_iscsilv - Creates a logical iSCSI Volume.
volume_iscsi_remove - Removes a logical iSCSI volume
volume_replication - Adds and removes replication to volume.
volume_status - Displays storage info.


FAILOVER:

Usage: failover --status | --actions | --start | --stop | --failover | --synchronize | --failback

This function allows you to stop, run or change the operation mode for the
given server.

Options:
-h, --help show this help message and exit
--status show iSCSI Failover status.
--actions show available iSCSI Failover actions.
--start start iSCSI Failover. On the primary server the Start
action
controls both servers. On the secondary server, it
controls
the secondary server only.
--stop stop iSCSI Failover. On the primary server the Stop action
controls both servers. On the secondary server, it
controls
the secondary server only.
--failover failover action delegate (switch) active server state
to the
passive server.
--synchronize synchronize data from the secondary/active server to the
primary server
--failback failback action will return the active server state to the
primary server.

A few more on iSCSI:

iscsi_target_assign - Assign lv with given name to existing iSCSI target.
iscsi_target_create - Creates a new iSCSI target.
iscsi_target_list - Lists iSCSI targets (syntax: alias;name).
iscsi_target_restart - Restart iSCSI target service.
iscsi_target_sessions - Shows and manages iSCSI target sessions.
iscsi_target_status - Lists the parameters of the selected target.
iscsi_target_unassign - Unassign from given iSCSI target lvname.