is it normal for md raid to rebuild upon bad shutdown?

[Red Squirrel]

My server locked up and I had to kill it with the power switch. When I tried to login from the console it just froze then beeped non stop.

It came back up and all looks ok but the raid is rebuilding. Is it normal, or should I be worried?

I checked /var/log/messages and did notice that one of my custom apps seg faulted shortly before the freeze so not sure if that caused some kind of kernel panic or something. It's a network app so it could of screwed up at the tcp stack level or something.

If a drive was failing would I get a decent indication somewhere? Like would the raid status tell me a drive is failed?

Right now all seems normal.

 

[Vadatajs]

I wouldn't say it's normal, but it's not inconceivable. I'd be more worried if it thought it had to rebuild on a clean reboot.

smartctl -a <drive> will give you an overview of drive health. There are also some tests it can do, more information in the man page.

 

[Nothinman]

All that means is that the md superblock was marked dirty so there probably was a write in progress when it hung so yes it should rebuild.

I checked /var/log/messages and did notice that one of my custom apps seg faulted shortly before the freeze so not sure if that caused some kind of kernel panic or something. It's a network app so it could of screwed up at the tcp stack level or something.

Unless your app was running in kernel mode it shouldn't be able to mess up the IP stack.