[j-nsp] M7i router with ATA disk going bad?

[sthaug at nethelp.no]

We have two different M7i routers that have each logged one (and only
one) instance of the following message:

atareadhdidentity: ioctl: Resource temporarily unavailable

and at the same time "show chassis hardware detail" (which is used by
rancid) shows that the ad1 disk is apparently absent. At the next rancid
run, everything is back to normal.

We have reported this to JTAC, and have been asked to run "smartd" to
perform disk testing and report the results back to Juniper. But we have
received no resolution to the problem (if it is indeed a problem).

Has anybody else seen this? We are somewhat worried that it could
indicate a disk going bad, since we have had a disk dying once on an
M10i (which of course uses the same RE and ATA disk as M7i). Note that
when the M10i disk died, we received no "atareadhdidentity" message.

Steinar Haug, Nethelp consulting, sthaug at nethelp.no