Extending the File system

[Govinda]

hi,

can u tell me whats the procedure and commands to extend file system in OVMS.

Thanks in advance
govind.

[Uwe Zessin]

You can add another (empty) disk to an existing one to form a so-called 'loosly coupled multi-volume' set. That's been available since a long time - I have first used in on VAX/VMS V4.x.

Starting with OpenVMS Alpha V7.3-2 you can grow the bitmap file and the addressable blocks through '$ SET VOLUME' when the underlying disk has been expanded.

[John Gillings]

govind,

>You can add another (empty) disk to an
>existing one to form a so-called
>'loosly coupled multi-volume' set.

Although Uwe is correct, please don't consider this option. Ancient and clumsy technology with many pitfalls and no real benefits.

I assume you're using "grow the file system" in the Unix sense - "add more disk space". The Unix concepts don't really apply in OpenVMS, as we deal with disk volumes independently of each other. You can add volumes to an OpenVMS system at any time, with no limits that anyone has enough money to reach.

The steps are:

1) Make the volumes available (exactly how depends on the technology)
2) Make OpenVMS see them (this may be automatic, or it may require executing SYSGEN IO AUTOCONFIGURE)
3) INITIALIZE to create a new file system on the disk or MOUNT to access an existing one.

Individual disks can be formed into arrays at controller and/or by the operating system. To make larger volumes, use RAID0 ("Striping"). To increase reliability use RAID1 ("Shadowing"). To get a bit of both (but with potential performance degradation) use RAID5.

When correctly configured, it is possible to expand disk volumes dynamically. This could be used, for example, to upgrade your storage from (say) 36GB drives to 72GB drives transparently. No downtime or interruption of service at all. The exact steps and commands required would depend on your storage technology. See the V7.3-2 New Features Manual for some examples.

Please be more specific about exactly what you're trying to achieve, what versions you're running and the type of storage you're using if you want more details.

[Mobeen_1 · Accepted solution]

Govind,
It really depends on how you have your disk subsystem implemented. I am not sure if you have raids, arrays etc.

If your disks are simple, then as Uwe mentioned add your disk (Sysgen IO Autoconfig), then init the disk and mount it.

A point to note here is ... if you are using the data/accessing disks using logicals, then you can mount this disk, edit the logical definition and include this disk....

The following is what i am trying to say

$ sh log MTDAT
MTDAT = DISK$DATA1:[000000]
= DISK$DATA2:[000000]

Now if you are adding another disk to have more disk space, modify that logical above to reflect your new disk DISK$DATA3

$sh log MTDAT
MTDAT = DISK$DATA1:[000000]
= DISK$DATA2:[000000]
= DISK$DATA3:[000000]

Regards
Mobeen

[Uwe Zessin]

Hello Mobeen,
I was not talking about a logical name search list pointing to several individual volumes - any creates will go to the first entry only.

I was talking about multivolume sets - see the help for MOUNT/BIND.

[Govinda]

THanks to al of u guys, i could able to do it.

Thanks.
Govind.

[Mobeen_1]

Uwe,
Sorry looks like there was some misunderstanding... i meant the following only with regard to your earlier statement. I know that you did not refer to the logical in your post


If your disks are simple, then as Uwe mentioned add your disk (Sysgen IO Autoconfig), then init the disk and mount it.


It was my suggestion that you can also grow the disk space using the logicals as i mentioned earlier. This was my thought and i have done this. As like any other approach this has its pros and cons. We used to have tough time with the converts process uding this approach.

I just meant to highlight another way of doing it.

regards
Mobeen

[Uwe Zessin]

OK, understood - thanks!