What happened to the PowerConnect 62xx v3.0 firmware?

Is anyone running this version yet?  Good/Bad? 

We can't afford to upgrade and then downgrade again (as happened with the last firmware upgrade).

Well just did the upgrade on a two 6248 that are stacked.

Upgrade went well, but with all the warnings in the upgrade procedure I was waiting for them to burst into flames.

Considering how involved the upgrade is, I would imagine a downgrade would be nasty or impossible.

Will be awhile before I can do any real stress testing.

Thanks.  Very brave!  :-)  I think we'll wait a few weeks/months to make sure that this upgrade is solid. 

I have two 6224 stacked switches that run my SAN and 1 6248 that is my network core.  Can't have problems with the switches or I'm out of business! :-0

The 3.2 bootcode CAN boot previous firmware versions.  2.2.0.3 bootcode cannot boot 3.2 firmware.   So, if you do decide to downgrade from 3.2 back to 2.2.0.3, it is not necessary to downgrade the bootcode.  So, if 2.2.0.3 is in your backup image, you would only have to change to boot backup image to downgrade.  Save your config file off the switch just to be safe. 

This is great news!

Master failover has been reduced, not eliminated.  The readme gives some percentages of the improvements.

As a test, with the simplest of configs, and nearly nothing else plugged into the swtiches except for another switch, I timed master failover at 8 seconds.  This was based on a conitnuous ping to the managment ip address of the stack.

management may take 8 seconds to failover, but datapath failover is sub 1 second in most cases.

Does that mean that when running this new firmware version, you will be able to upgrade one switch at time to a future firmware version w/o taking the whole stack down (or at least taking it down for just a second)?

We're planning to use two PowerConnect 6248 switches for a HA setup and we cannot afford both of them to go down at the same time for more than a second or so.

If the new firmware still doesn't allow to update one switch at time for future firmware versions, then we're going to use a LAG between the two switches and manage them separately instead of the stacking.

Thanks.

No.  This firmware does not provide hitless upgrade of a stack.

The reason is that switches must be running the same firmware in order to link into a stack.  So you cannot update a single switch in a stack, because it will not rejoin the stack because of the firmware mismatch.

The improvement in this release is non-stop forwarding.   This means if your master switch went down or was taken down by an administrator, the backup master would take over switching and routing quickly (sub 1 second in most cases).

Since we're trying to achieve a truly HA setup for our VMware cluster (each ESX host will have a connection to one switch and another to the second switch) and we want to keep the possibility to upgrade to future firmware releases for the switches with no downtime for the VMware cluster (by upgrading the firmware for one switch at time), it looks like it's better for our setup to not use the switches stacking feature, but instead use a LAG (e.g. dual 10GbE) between the two switches and manage them independently, do you agree?

Besides a little bandwidth, I don't see losing anything else by not using the stacking feature.

Anything else (e.g. special config, spanning tree setup, etc.) we should be aware of when interconnecting the two switches with the LAG instead of the stacking?

Thanks again.

If you need to do a hitless upgrade, do not use stacking for redundancy.

Stacking provides the increase in bandwidth as you mentioned, and a single point of management.   So you would lose the single point of management. 

There are no special configuration concerns when using a LAG instead of stacking. 

Firmware 3.2.0.7 has been out for a little while now.  I have been running it for weeks now and it seems to be stable.

http://support.dell.com/support/downloads/download.aspx?c=us&l=en&s=gen&releaseid=R269990&SystemID=PWC_6248&servicetag=&os=WNET&osl=en&deviceid=12748&devlib=0&typecnt=0&vercnt=9&catid=-1&impid=-1&formatcnt=0&libid=5&typeid=-1&dateid=-1&formatid=-1&source=-1&fileid=397692

3.2.0.9 has now been released and, from the release notes, appears to fix a lot of bugs!  I might be brave enough to try it now.