[Info-vax] Transient anal/disk errors

[tadamsmar]

On Thursday, December 12, 2013 1:51:43 PM UTC-5, Stephen Hoffman wrote:
> On 2013-12-12 17:05:38 +0000, tadamsmar said:
> 
> 
> 
> > If I dismount from a shadowset and run anal/disk on the dismounted 
> 
> > disk, I sometimes get warnings form anal/disk.
> 
> 
> 
> If the applications are not quiesced or shut down when the shadowset 
> 
> member is removed from the shadowset, then some errors and 
> 
> inconsistencies can arise.
> 
> 
> 
> > I just recently started checking this as I have gotten a bit more 
> 
> > concerned after I had persistent warnings from anal/disk on one system 
> 
> > that required a good bit of cleanup.
> 
> 
> 
> I'd start by replacing the disks, then a combination of wholesale 
> 
> server replacement, and a software upgrade.
>
>If you have the 
> 
> prerequisites available, probably with Itanium.

Prerequisites being willingness to pay a bunch to go from one dead end
legacy system to another.


> You're running on 
> 
> hardware that I wouldn't recommend to most hobbyists, and you're 
> 
> spending far too much time keeping it going based on the various 
> 
> postings with the same problems over and over again,

We have had few problems and have spent little time on them overall.

> which usually 
> 
> means that the technical and IT discussions involved here are secondary 
> 
> to some non-technical or managerial issues (usually funding, but there 
> 
> can be other triggers), and that's not a fun place to be when you're 
> 
> trying to keep fossil-grade servers available and operational.
> 
> 
> 
> > So perhaps even a dismount from a shadowset is a bit risky for backing 
> 
> > up.  Of course, it's a heck of a lot easier than the official way to 
> 
> > prep a disk for backup.
> 
> 
> 
> If your disks are dismounted when the applications are quiesced, and 
> 
> you are getting corruptions, then some combination of hardware and 
> 
> software are failing you.

I am not sure I had corruptions under those circumstances.

The persistent ANAL/DISK warnings arose after I had shifted the disks around to
two systems, something might have gone wrong then.  Anyway, I am treating it like one-off special case but I am monitoring for problems.   And I am doing a bunch of stuff to better manage one off special cases, and to minimize the need to move disk around.

I bought 3 bare-bones systems to put in storage as backups, that will mitigate the need to touch other running systems (to use them as backups) if one system fails.

Also, I am establishing better manual and automatic procedures to check disks for errors and ANAL/DISK warnings during any procedure that involves breaking up a shadowsets and moving physical disks.  I do a daily routine check for errors, but I need to check repeatedly at various points when I have to do something atypical with a disk.

At least, if I had been more careful and did more checking while I was making changes I would know exactly when the persistent ANAL/DISK warnings arose.


> 
> 
> 
> 
> 
> -- 
> 
> Pure Personal Opinion | HoffmanLabs LLC