We had a power outage in our DC, and after that one of the EX3300 in VC can't start. The switch was running normally before power loss. Right now it's in boot loop. I appended the log from console. It tries to boot from disk0s1 and disk0s2, but they throwing a lot of errors.
What should I do now? Is it an hardware issue? Or just software?
Very good question mate! this could be due to many reasons, A software/application/process could hang while attempting to open a file, data could get damaged while copying a file to a flash drive, an abrupt power loss could disrupt a save action, etc.
How to avoid it? very good question too, well these causes I mentioned are all fairly normal functions that could lead to data damage if some form of disruption occurs hence preventing file corruption starts with proper device care e.g.
Shut down the system properly and in case of power outages have a redundant power source and PSU's if applicable/possible, don’t delete system files readily ( if you need to delete something make sure it is not a file that is paramount for the switch integrity and vitality, check hard drive regularly and a general health check on the hardware/software is very.important, and my best advise is backing up an installation using Snapshots: https://www.juniper.net/documentation/en_US/junos/topics/topic-map/backing-up-install-using-snapshot...
The main protection from power failures on ex switches is to use snapshots. These create a backup boot partition that can be used if an event corrupts to boot partition. These do need to be keep up to date with any changes in the Junos upgrade and configuration.
When a corruption occurs you boot from the backup partition and a message is show on the cli and trap created to alert the operator that the original partition is corrupted and you need to run snapshot again for the repair.
Steve Puluka BSEET - Juniper Ambassador IP Architect - DQE Communications Pittsburgh, PA (Metro Ethernet & ISP) http://puluka.com/home