access beyond end of device
03:00 rw=3 want=1715620 limit=1714018
What's strange here is that the device is 0x3451E0 sectors in size, but
Linux starts to hurl big chunks while the mirroring tables are being
opened on a device. Under NWFS 2.0, there is an "open" bit in the
mirror tables that gets written to identify they have been mounted to
the mirroring manager (no volumes have been mounted at this point, the
driver is simply loading). This is how we detect in-sync partitions
during power failures, etc. since we know that if someone was left open,
they are probably not in-sync (though they may claim to be in-sync).
When we shutdown the driver, partition open bits are cleared and written
to the mirroring tables to identify any partitions that in in-sync at
shutdown.
This error is claiming that sector 1715620 (0x1A2762) is beyond the end
of the device, which in fact it is not. We got it to go away by forcing
the first open of all the mirrored partitions after the system reaches
run level 3 (the first time someone tries to mount a volume), instead of
doing it when the driver loads.
I am assuming this may be happening becuase while the /root fs is
active, the range of disk access is somehow being restricted, though I
am guessing here. Are we doing something bad that could screw something
up? Is this error a bug or should we hold off on opening the mirrored
partitions when we get to run level 3 but prior to fstab auto-mounting
of volumes?
Please Advise,
Jeff
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/