Zoning Issue

[RotorConsultant]

Ok we have a strange issue. We are continually having VMWARE ESX hosts drop connectivity.  It seems to only affect one fabric.  The SAN consists of a HP 3par V800, connected to 48K switch and IBM chassis 4020 switches.  Nothing has changed on the fabric and just started happening about two weeks ago.  The odd thing is there are no errors and the ports show up and operational.  The only way that it will resolve is to inactive the zones for that server (doing 1:1 zones) and readd and reactivate them.  Again only affecting VMWARE hosts.  We figured it was a VMWARE bug but the odd thing is that is only affects one fabric and the fix is to unzone and rezone the links.  I have a support case open but figured I would throw it out here as well.

[Guldmyr]

"The odd thing is there are no errors"

Did you look in the port errors of all the ports in that fabric? (output of porterrshow might help you make a delta).

[ravi.shankar3]

Did you get the solution from support , what was the error.

[RotorConsultant]

A cable to he blade switch had become loose. We had been watching for CRCs and ITWs and finally got a break when a port spiked with high errors which was making the whole fabric unstable. We went to test the cable and it pulled right out.  We haven't seen any issues since.