Re: mkfs.ext3 taking long time on raid1 with nda disk

"Also sprach Michael Rendell:"
>   Have noticed that creating a file system on a raid1 device
> that uses a enbd device sometimes runs into problems: looks

Which enbd? You are using enbd 2.4.35 aren't you? That's the stable
version.

I've also just noticed that since about kernel 2.6.12, the kernel's
32 bit jiffies rollover happens 5 minutes after boot! That probably
causes huge instability for the first five minutes in enbd, since
most timeouts will be the other side of int_max from start time. I've
rapidly done a hardening against rollovers myself. 

You sound like you're efficient enough to get things working within
five minutes!

> like it is geting errors writing to the server and taking a
> while to recover.  Wondering if something is wrong in the
> configuration?

Nothing can be wrong :).


>  # raid now created; mkfs running...
> 
>     Dec 17 17:18:01 nmhd-bs1 kernel: ENBD #1629[0]: enbd_rollback (0): 
> erroring too old (delay 18s >= timeout 90s) req cb2f9078 id 0x4745ca1f!


Yes, this is a symptom of having crossed the 32 bit jiffies rollover boundary.

It is not normally the case that 18 is greater than 90!

And I recall splitting that message up into a+b>c (the above is c-a>b).

Wait 5 min after booting for the mo. I'm on this.

Peter