dropped frames, Dynamic Load sharing and data corruption

[Mossman]

Hi

I have a customer who has asked:

"Do dropped frames signify unrecoverable data loss".

They have a fully HP hardware environment (EVA6000, P-Class blades, HP Brocade switches etc), are running VMware and are experiencing unstable hosts which are aligned to 1 particular vDisk and 1 particular server in the BladeSystem.

VMWare have suggested that it may have something to do with Dynamic Load sharing on the Switches but aren't 100% sure.

Are you able to shed any light?

I look forward to hearing from you.

Best regards

Richard

[ben horan]

Having DLS disabled on the switches will mean that the switches route traffic via fixed routes rather than deciding the route on a frame by frame basis. This frame by frame determination may increase the amount of dropped frames.

Do you have Continuous Access?? The best practises are for DLS to be disabled if you use CA on your EVA.

[Graeme Wood]

Hi,

Does anyone have a document that shows the recommended setting for CA?

Cheers,

[ben horan]

Here's a link to best practises with CA and Brocade B-series switches:

http://h20000.www2.hp.com/bc/docs/support/SupportManual/c01081582/c01081582.pdf

[Graeme Wood]

Thanks Ben,

HP support had recommended it but could not show me it in writing in a document.

Cheers,