JimNim's questions

Does Windows Storage Spaces implement optimized/reduced rebuilding of a mirror or parity drive group if a drive is mistakenly removed and added again?

I have some limited exposure to storage spaces in an enterprise environment, and toy around with it a bit at home as well. I noticed recently that when accidentally disconnecting a drive that is part of a mirror, then reconnecting it shortly after (15-30min), it seems like the time spent in a "Repairing" state is drastically lower than if a drive is removed and "replaced" triggering a full re-mirror.

I can't seem to spot any write-ups or documentation on functionality like this, but I could imagine that Storage Spaces would perform some kind of tracking on which "slabs" are modified during the time the offending drive is not connected, and only performs a repair on the applicable data.

Any first-hand experience or observations would be helpful, though documentation would be helpful (though I'm fairly sure there is none specific to this after a good amount of searching).

How can the total amount of port buffer memory be determined for Cisco switches?

Cisco doesn't seem to publish this data anywhere as part of the hardware specifications, but the information seems crucial for determining whether or not a switch is adequate for iSCSI SAN traffic.

I see various values shown in the output of the show buffers command, but I'm not quite understanding how to turn this into usable data for comparing switches.

I'm particularly interested in comparing the Catalyst 2960 and 3750 models, specifically to help deter someone from using an inadequate switch for iSCSI traffic. Any thoughts on other factors that play into switch usability are welcomed, though I'm particularly interested in port buffer space data here.