Hi,

I have two Sun Fire V490 with Solaris 10 5/08, FC switch and two Sun StorageTek 3500 FC arrays. Each array is connected to switch and to one server at a time. In the last week I installed Solaris 10 5/08 on both servers and set up Sun Cluster, version 3.2. At wednesday all was fine - all arrays were visible to all servers, at thursday I didn't work with them, but at friday, when I begin to work, I was unable to access some of them from both servers. On both servers I have two QLogic HBA, I don't know exact model now. Unfortunately, I can't provide more output, because I can access them only tomorrow, but I want to know - what can be the cause to such odd behavior - two days arrays are visible, at third - not?.. What output should I provide to troubleshoot this?..

have a look at the commands "cfgadm", "luxadm" and "fcinfo".

Thanks for your answer! I know about these commands and my question is why if arrays are visible at one day, they are not at next? How can I determine the source of such strange behavior?.. I can't post outputs of commands you provided, but I remember that fcinfo hba-port -l showed loss of sync and signal count equal to 1 at last time I performed it. Moreover, after loss of visibilty for OS, probe-scsi-all sometime buzzed when triyng to discover peripheral hardware...

sounds like a SAN problem...

What SAN if I didn't configured it? Maybe I should?.. Tomorrow I'll be onsite and collect as much information as I can...

Recently, I've found that our switch is actually for SAN so my problem seemed to be incorrect SAN configuration. Also my colleague told me that they perform some initial configuration of this switch (Brocade 200E)... I didn't know about. I think it extremely changes an approach to solve my problem...

as i told you, it sounds like a SAN problem