Originally posted by: Seby
Originally posted by: NBS
Originally posted by: mackem
Seby,Im pretty sure all versions of MC on the USP allow the option of performing the MC exchange with the ports online. Of course this procedure is only supported when done by an HDS certified engineer. You will probably want your engineer to use the non-stop SCSI mode. They will know what you are talking about - if they dont then ask them to leave site and get another engineer in ;-)Nigel
Originally posted by: Ted
Originally posted by: Ayix's
Yes Ive done a couple with the non stop SCSI option (well actually Ive planned and supervised them but an HDS engineer has actually stood at the SVP.....).Ive not had any problems and your hosts will not fail over paths because the paths will not go offline. You do have to make sure that you do the work at a time when the MPs/CHIPs are not so busy. I "think" (from the top of my head) the process has some pre-requs including MP useage be below 50%. This is because as each processor is rebooted its associated ports will be moved to another processor with the potential of doubling the workload for those processors that are not rebooting.Im certain your HDS engineer will have experience doing this operation.
Originally posted by: Chris Cade
Make sure your HDS engineer doesn't overlook the default HDLM setting that causes paths to not return to normal state after a microcode update (AutoFailback)For some reason auto failback is off by default in HDLM and your hosts will go down after the second controller is updated if this setting isn't on/opt/DynamicLinkManager/bin/dlnkmgr set -afb onThis should really be on by default, but its not. You can also view your current config by doing:/opt/DynamicLinkManager view -sys -sfunc
Retrieving data ...