[Info-vax] Raid 5 Add Disk to Raid 5 with p400 and MSA$UTIL

[Neil Rieck]

On Tuesday, April 4, 2023 at 8:12:08 PM UTC-4, Russell Allred wrote:
> On Tuesday, April 4, 2023 at 4:36:25 PM UTC-7, gah4 wrote: 
> > On Tuesday, April 4, 2023 at 4:04:59 PM UTC-7, Arne Vajhøj wrote: 
> > 
> > (snip) 
> > > or the RAID controller can actually handle it automatically I don't 
> > > know. Someone that know that RAID controller may be able to tell. 
> > > 
> > > But just to understand the basics: 
> > > 
> > > old system: 
> > > 
> > > disk1 disk2 disk3 
> > > blk 1 blk 2 blk 1 xor 2 
> > > blk 3 blk 3 xor 4 blk 4 
> > > blk 5 xor 6 blk 5 blk 6 
> > > 
> > > new system: 
> > > 
> > > disk1 disk2 disk3 disk4 
> > > blk 1 blk 2 blk 3 blk 1 xor 2 xor 3 
> > > blk 4 blk 5 blk 4 xor 5 xor 6 blk 6 
> > Some controller (or computer) might be able to do that. 
> > 
> > It isn't obvious that an intermediate state is good, though, 
> > so you probably want a full backup before starting it. 
> > 
> > Seems to me that with appropriate checkpoint, so that it 
> > remembers how for it has gone, and how far still to go, 
> > that it words. And likely when no-one else is writing. 
> > 
> > You can generate the whole new contents of disk4, 
> > without changing the other ones. Some disk space, 
> > somewhere else, for temporary use would be nice. 
> > Copy some blocks to temporary space, record that 
> > you did that somewhere, generate new blocks for 
> > those parts of the disks. Repeat until done. 
> > 
> > If it fails partway through, continue on based on 
> > the known stored temporary blocks.
> I get the theory part ok. It's can the MSA$UTIL program and my P400 controller do it? If so, what are the commands? I have found this link which helps: 
> 
> http://h41379.www4.hpe.com/doc/84final/6048/6048pro_052.html#bottom_052 
> 
> I hope someone has tried this and can show me examples of the commands to use and pitfalls to avoid.

I'm just running out the door so do not know if these examples meet your needs. But they might be worth a look

https://neilrieck.net/docs/openvms_notes_itanium_diary.html#msa

OpenVMS systems can run unattended for years which means that companies forget about them until they run into trouble (like running out of storage space) so selective purging might help you in the sort term until you decide what to do next.

For example, you might wish to consider JUST the first tip on this page:
https://neilrieck.net/docs/openvms_notes_system_manager.html

In an emergency you might wish to purge all the log files everywhere.
Next, consider purging all the datafiles everywhere.
I assisted on an emergency last year where where I was required to purge everything (but I did a "keep 3" just to be safe)

Neil Rieck
Waterloo, Ontario, Canada.
http://neilrieck.net