[Info-vax] Are queue manager updates written to disk immediately ?

[Simon Clubley]

On 2013-04-19, Paul Sture <nospam at sture.ch> wrote:
> Another thought occurred to me today.
>
> Was volume shadowing in operation on the disk that holds the queue 
> manager database?
>

No. Multiple hardware RAID 1 sets all connected to the same controller,
no writeback caching in operation (everything is write-through), no disk
failures, no filesystem corruption (ie: from disks been out of sync).

That last item is critical. If the disks were out of sync you would
see filesystem corruption been reported by analyze/disk because the
disk for read I/O in a RAID 1 set is chosen effectively at random.

No application level detected corruptions either.

> I am thinking along the lines of the "wrong" disk coming back as the 
> shadow master here.
>
> What caused the power loss?
>

Something external to the building containing the machine.

Simon.

-- 
Simon Clubley, clubley at remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world