04-13-2005 10:33 AM
Just received a Sun alert (57687) stating that the 3850 switches could panic after uptime of 497 days or more. The fix is to reboot or upgrade to 4.2.2. My 3950's are in a fabric with 3800's which can only upgrade to 3.1.2 (as told by the reseller). I was told it is recommended to keep the 3850 to the same OS level. Is the reboot the only solution and does this affect the 3800/3200 family of switches.
04-15-2005 05:29 AM
The 3850 is running a 4.x version of Fabric OS and can't run a 3.x version. You should be fine upgrading the 3850 fabric OS as long as the release notes for 4.2.2 say that a 3850 can connect to a 3800 running the 3.x version you are running. You may need to upgrade the 3800s to a newer level to be able to connect to a 4.2.2 3850.<!
04-06-2006 10:34 AM
Just so you know, the fix you speak of is in v4.2.2b, not v4.2.2. Check the release notes for v4.2.2b, but I do believe that v3.1.2 is OK with v4.2.2b. Also I would ask the reseller why can't you go to 3.2.x on your 3800's. The 497 day bug only affects linux based switches (12K, 24K, 3900, 3850, and 3250).
04-06-2006 11:39 AM
I too found out the hard way. One of my EMC branded 12K's rebooted. I escalated a call to see why and when it might happen again. The CE said, "I can tell you exactly when it'll do it again, 497.1 days from now."