i was upgrading my 3560G-24PS when one of them failed and would keep booting and then restarting , i have upgraded another exact same switch in the same rack with no issue but this one wont boot successfully from the new ios , i have to use rommon to boot from the old one in order to get it online
i have attached what is appearing on the console when it failed to boot
The output doesn't look good stating memory corruption but its strange then you can still boot it back to the old ios if it was fully hardware failure
Have you confirmed the IOS that you uploaded didn't get corrupted somehow and the md5 hash is the same as the website
if you get a fresh copy ios and its still the same or you try another version like SE9 if its still the same something may have gone in hardware alright and you may need to RMA it if still supported
if the md5 is ok the ios is probably ok , all ios are 128/16 in dram and flash for that platform so any image should work on it really theres no specific jump versions required on this series, as a last resort you could try SE9 its an main deployment release or something else but if its not booting SE10 and the md5 is ok it probably wont work either unfortunately but maybe worth a shot
the dump is not good as it showing memory failure and a trackback outputs , if its not able to upgrade its software something's gone in hardware only TAC will be able to say what exactly but they will probably just recommend an RMA
i tried yesterday to upgrade to "12.2.35-SE5" then to c3560-ipbasek9-mz.122-55.SE10.bin , it worked but just as the switch booted it initiated a restart after it went into the cli normally , i thought that was it , but i found when i came today it reverted to the old ios , it is showing this under show ver
System returned to ROM by error - Debug Exception (Could be NULL pointer dereference) Exception (0x2000)!, PC 0x1BA213C
what does this means?
should i try to upgrade again to 12.2.35-SE5 then to a previous version like 122-55.SE9?
It could be this below but your already on the higher image , you could try SE9 see if it will stay stable with that , its risky but you could remove the old image so it cant go back to it force SE9 or another IOS , but then you could put the switch into rommon if it fails on SE9 as it will have no fallback image
Symptom: A Catalyst 3560 reloads with the following error message: Debug Exception (Could be NULL pointer dereference) Exception (0x2000)! Conditions: The Catalyst 3560 switch is running Cisco IOS 12.2(20)SE4 (and later) and PD is detected/classified as "overcurrent class". This problem won't happen with latest images starting from 12.2(25)SEA