MPX110 replication

 
Kehad Snydewel
Frequent Advisor

MPX110 replication

Hi Guru’s

I have 1 MPX110 per site with a 5Mb/s ISL. This is a shared link for Replication and network traffic but it’s basically a new implementation and no network traffic is currently on the ISL.

HW Config
PRD SITE
EVA6000 (XCS6.220)
2x 4/32 B-Series SAN Switches (FW=v6.4.0a)
1x MPX110 (FW=2.4.4.1)

Dr Site
EVA4400 (XCS09534000) with embedded switches (FWv6.4.0a)
1x MPX110 (FW=2.4.4.1)


I managed to setup the FCIP link and can see both EVAs in CV. I also created a test DR Group (Async mode) with a 10GB vdisk which replicated successfully from the EVA6000 to the EVA4400.

I would like to create a new DR Group but it fails with error “cannot find storage system”. I forgot the actual error message but it is something like this.

But I can add a vdisk to the existing test DR Group…..

Zoning is not configured on the Switches but I will configure zoning soon.

What is the maximum number of DR Groups that I can configure when using a shared 5Mb/s ISL or does it depend on the amount of data, amount of changes, etc., etc.?

How can i test the ISL's performance?
5 REPLIES 5
Eemans Dany
Honored Contributor

Re: MPX110 replication

Hi,

Can your MPX110 see all the EVA ports (4 one side and 4 on the other side).

If not you need to configure the DR to use specific ports (visible on the MPX on both sides).

Dany
Kehad Snydewel
Frequent Advisor

Re: MPX110 replication

i only have 1 fabric (Fabric A) connected so i only see FP1 of both controllers (2 connections/ paths)
James Coppock_1
New Member

Re: MPX110 replication

Hi

I had a similar issue with 8400 to 6000 CA. The cause was basically down to Exchnage Based Routing.

In the 09534000 XCS code an option has been added to modify the data replication protocol, it defaults to 'HP SCSI FC Compliant Data Replication Protocol' or EBR. Unfortunately this is not compliant with Gen 1 or 2 EVA's, for me it manifested itself in Command View as 'no paths' but when you tried to create a DR group the other array was visible and it's Disk Groups could be seen, if you then tried to create a DR Group you get an 'unable to connect' type error.

I changed the option to 'either' and my paths came online. I did however then have errors in the controller event logs complaining of 'excessive data exchange retry rate' during replication (Synchronous). I have not been able to make the change yet but it is expected that this is being casued by Exchange Based Routing and DLS being enabled in the fabrics (4/32 and 8/80). I will be disabling EBR and DLS and setting Port Based Routing.

CA on EVA prior to 09534000 is only supported with Port Based Routing and no DLS.

You need to do switchdisable to make the fabric changes.

The EBR option in Command View (only on the 09534000 array) is under 'Data Replication', 'Data Replication Options', 'set Data Replication Protocol'.

Kehad Snydewel
Frequent Advisor

Re: MPX110 replication

thansk guys,

i had to configure GE2 and FC2 on MPX's. i can add more then 1 DG Group now.
Kehad Snydewel
Frequent Advisor

Re: MPX110 replication


configure all GE and FC ports on MPX's and play arround with the port speeds on the SAN switches as well.