Backup .com to look at more than one disk

[grwere]

At the momment backup .COM only looks at dsk01. We would like it to look at both disk areas. BUT - we also don’t know if any other db jobs look at all of the snp, rda files too. Getting the TDMDBF com file look at dsk01 AND dsk02 for it's rda & snp files is probably quickest.

Thamks in advance.

[Shriniketan Bhagwat]

Hi,

>>Backup .com to look at more than one disk
Do you mean to take BACKUP of files from multiple disks from the singe BACKUP command?

In this case you can do that by using /INPUT_FILES qualifier. Here you need to specify the files name (which you want to BACKUP) from multiple disks in a file and that file needs to be specified as the input specifier in the BACKUP command.

Regards,
Ketan

[Steven Schweda]

> At the momment backup .COM only looks at
> dsk01. We would like it to look at both
> disk areas.

My psychic powers are too weak to tell me
what this system is, or what "dsk01" and
"dsk02" are, or what's in your "backup .COM".

> BUT - we also don't know [..]

And you think that someone here _does_ know?

> [...] the TDMDBF com file [...]

Yet another mystery.

[Shriniketan Bhagwat]

Hi,

Yes, as Steven said, you query is not so clear. Please describe.
Below is the small example how you can backup the files from multiple disks.

$ dir VISHAK$DKA100:[000000]TEST.COM;1

Directory VISHAK$DKA100:[000000]

TEST.COM;1

Total of 1 file.
$ dir VISHAK$DKB200:[000000]WAIT.COM;1

Directory VISHAK$DKB200:[000000]

WAIT.COM;1

Total of 1 file.
$
$ type backup.com
VISHAK$DKA100:[000000]TEST.COM;1
VISHAK$DKB200:[000000]WAIT.COM;1
$
$
$backup/log backup.com/input_files temp.bck/save
%BACKUP-S-COPIED, copied VISHAK$DKA100:[000000]TEST.COM;1
%BACKUP-S-COPIED, copied VISHAK$DKB200:[000000]WAIT.COM;1
$
$
$backup/list temp.bck/save
Listing of save set(s)

Save set: TEMP.BCK
Written by: KETAN
UIC: [000200,000006]
Date: 29-JUN-2010 20:14:41.12
Command: BACKUP/LOG BACKUP.COM/INPUT_FILES TEMP.BCK/SAVE
Operating system: hp OpenVMS I64 version V8.3
BACKUP version: V8.3
CPU ID register: 80000000
Node name: _VISHAK::
Written on: _VISHAK$DKA100:
Block size: 32256
Group size: 10
Buffer count: 403

[000000]TEST.COM;1 1 22-APR-2008 11:57
[000000]WAIT.COM;1 1 19-JUN-2010 03:36

Total of 2 files, 2 blocks
End of save set

$
$

Regards,
Ketan

[Shriniketan Bhagwat]

Hi,

Welcome to ITRC. Please provide more explanation when asking the questions, So that you can get better answers. :-)


Regards,
Ketan

[grwere]

Thanks for the advise, I will get more info as it becomes available. But will give the suggestions already listed a try for now.

Thanks again

[Joseph Huber_1]

If You try it and detect the used VMS/backup version is too old (before 7.3-2 or 8.3 whenever the /INPUT qualifier was introduced),
depending on the length of the list, logical names and logical name lists can be used to achieve the desired result.

Example:
define todo disk1:[dir1...],disk2:[dir2...]
backup todo output.bck/save

But be aware if there are the same directory names on different disks: on restore they can't be directed to different disks, they will be mixed into the same disk;[dir...]: which may be wanted or not.

[P Muralidhar Kini]

Hi,

Welcome to the ITRC Forum.

For more information on OpenVMS Backup, refer
http://h30266.www3.hp.com/odl/vax/opsys/vmsos73/vmsos73/6017/6017pro_044.html
- Formulating a BACKUP strategy
- Setting process quotas for efficient backups
- Using disks and tapes
- Listing the contents of a BACKUP save set
- Backing up user disks and volume shadow sets
- Restoring user disks and volume shadow sets
- Backing up and restoring the system disk
- Ensuring data integrity Section
- Troubleshooting Section

This talks about how you go about using the OpenVMS BACKUP.

Hope this helps.

Also, as you are new to the forum,
Refer the following link which says how you can thank the forum -
http://forums13.itrc.hp.com/service/forums/helptips.do?#28

Regards,
Murali

[P Muralidhar Kini]

Hi,

Some more documents -

* HP OpenVMS System ManagementUtilities Reference Manual: Aâ L
http://h71000.www7.hp.com/doc/83final/6048/ovms_83_sysman_util1.pdf
-> Section - "7 Backup Utility"

* The OpenVMS Frequently Asked Questions (FAQ)
http://h71000.www7.hp.com/faq/vmsfaq_011.html

Regards,
Murali

[Shriniketan Bhagwat]

Hi,

Files with .RDA and .SNP types are basically RDB files. You can use RMU to backup such files. Refer the below link on RMU reference manual.

http://www.openvms.ru/rdb$doc/oraclerdb/rmurm7/rmu_ref_pro_rmucommands.html


Regards,
Ketan

[Hein van den Heuvel]

grwere,

Review your query. All the specifics mean very little to most readers as they only have a meaning in your context.
So open up ( $ EDIT/READ or $ TYPE ) the command files ( backup.com? tdmdbf.com?) being used an look for common terms we can all relate to. Specifically BACKUP and/or RMU commands.

Ketan makes a good point about RDA and SNP files typically refering to RDB databases.
For those RMU BACKUP is much preferred over OpenVMS BACKUP.
RMU knows where all related datafiles are, whatever disk (dsk01? dsk02?) they are on. It'll figure that out. No worry for you.

A better reference for RMU is:
http://download.oracle.com/otn_hosted_doc/rdb/pdf/rmu_ref_pro.pdf

As you prepare your detailed question (if still needed), be sure to indicate with OS (OpenVMS? :-), Database (RDB?), Platform ( Alpha?), Storage setup ( eva? ) and verision numbers aas available.

Good luck!
Hein.

[abrsvc]

I'm restating was has been said to a degree here, but for claification purposes...

1) If you are attempting to backup a database that spans multiple disks, then the RMU Backup is the correct method to use. This way the entire database is saved corectly along with the component distribution.

2) VMS Backup can then be used to backup the "backup" to tape. RMU backup will create an .RBF file (RDB Backup File).

3) The RMU Backup will also retain all of the snapshot status.

Dan

[Andy Bustamante]

Assuming you have an Oracle RDB database,
unless the Oracle RDB database is shut down, BACKUP will not be able to create a useful copy of the database files. You need to use the RMU utility to preserve your database files. RMU can write to disk or tape. If writing to disk, VMS backup can create backups of the database backup.

Backups are worth having, but it's restores that are useful. It sounds like something may have changed in your enviroment and the backup never updated. Schedule time to test your backup strategy once you've made any changes.

[dabney121]

Hello, I work with grwere. The issue is that we are out of disk space and one of our processes wont run. ALL of the .snp & .rda RDB files have been located on dsk01. When there was not enough space for some of the files to expand, we tried moving some files to another area. That is also when we discovered that the TDMDBF disk to disk backup "looks" at ALL of the files and expects them to be on dsk01.
We would like to define a logical so that we can have it look to another disk but are not sure how to define it. Probably simple for you guys but we are not experienced admins for this type of system.

We have a very old version of Openvms 7.2-2.

[Hoff]

Why not drop fifty or a hundred gigabytes of (gently used) SCSI disk storage into this OpenVMS Alpha V7.2-2 configuration? Or more?

That's a simple and effective and inexpensive solution for many of these cases, even if you have to add a disk shelf and upgrade a SCSI controller in the process.

[Steven Schweda]

> [...] That is also when we discovered that
> the TDMDBF disk to disk backup "looks" at
> ALL of the files and expects them to be on
> dsk01. [...]

You have the advantage of being able to see
what this "TDMDBF" does. The non-psychics
here can't see it.

> We would like to define a logical [...]

And perhaps you can. And perhaps someone who
could see what "TDMDBF" does could suggest
something.


> Why not drop [...]

A fair question.

[Andy Bustamante]

As Hoff points out, adding storage may be the workable solution. Be cautious about simply copying files from disk to disk. First, you need to let RDB know about the migration, second, depending on disk use it may not be an appropriate allocation.

For example, a typical configuration would have data files on disk_A and journal files on disk_B, with a regular RMU backup copied to disk_B. Disk_B could also be designated for temporary report file use, which would generate much I/O and create potential for fragmentation if mixed with database files.

If no one at your business has a the knowledge to reconfigure this system and avoid creating additional long term issues, consider retaining a consultant. There are several of those on available on the forum.

Migrating to a supported version of the operating system (7.3-2 minimum or 8.3 current) and supported version of RDB should also be considered.

[dabney121]

Can you define 1 logical for 2 disks so that our process thinks it is 1 disk?

[Steven Schweda]

> Can you define 1 logical for 2 disks so
> that our process thinks it is 1 disk?

Define "our process".

It's hard to make two disks appear exactly
equivalent to one disk. It can be easier to
do some other things which might be useful,
depending on exactly what you're doing.

[Hein van den Heuvel]

>> we are not experienced admins for this type of system.

That's ok, we are. :-). But we do nor know diddly squat about your setup, although we've been know to speculate with the best of them.

Anyway, you can read right? So go slowly over the earlier replies. Focus on the RMU versus BACKUP. If is very possibly that your 'TDMDBF' has been doing a fine job or has not been doing anything useful at all (for years).

If your company did not hire the right folks with the right experiences for the job, then they may need to hire some outside support to bring you guys up to speed, and make sure the system is in good shape were it can recover from trouble.

>> we discovered that the TDMDBF disk to disk backup "looks"

We have no clue what that thing might do or not do. Help us help you by explaining what it seems to do! In questioning that you might just figure out what it really does.

>> The issue is that we are out of disk space and one of our processes wont run

So that was fixed by adding a 'disk' and now you don't know how to backup the resulting mess?

Can the failing process use its (sys$scratch?) space elsewhere?

>> Can you define 1 logical for 2 disks so that our process thinks it is 1 disk?

Yes, in several ways which may or might not help you.
Possible solution are:
1) mounting them as a bound volume set.
(Go slowly on that. The way back is tricky!)
2) putting directories on multiple disks in search lists.
3) upgrade to OpenVMS 7.3-2 and hardware to grow / combine the disks, followed by Dynamic Volume Expansion (DVE) ( SET VOLUME/LIMIT ) to make it accesible by OpenVMS

Good luck!
Hein

[Jan van den Ende]

grwere,

Let me (too) start with welcoming you!

If the functioning of this system has some value to your company, I would really, I mean REALLY, advise the company to get some external help!

Just ask your management what would happen if this system SUDDENLY stops ALL functionality.

If that can be estimated in money, consider hiring a seasoned consultant as insurance costs.

From your profile, you are located in USA.

Surely there are many others, but just from the top VMS members Hoff, Hein, and Robert are USA based, and offer such services.
I know all 3 of them to be VERY good at these things, and they also are seasoned trainers that will pretty quickly bring you at a sufficient level of knowledge of YOUR system.

Being located in Europe myself, I do not count myself as a likely choice, but I can be reached at JAN at dot NL

GET EXPIERENCED HELP, before it is too late.

Proost.

Have one on me.

jpe

[abrsvc]

If you are located in New England, shoot me an Email. I can help with this. It would be easier to explain things in person rather than a series of notes in this thread. There are a number of possible solutions to the problem. Each has its pros and cons.

Dan

[Bill Pedersen]

grwere and dabney121:

As Dan mentioned, sometimes it is useful for face to face discussions to solve the problem and give a bit better insight to the solution.

Many of us are located in various locations in the US, if you are so located and would generally be happy to give more specific assistance if necessary. Of course there are others here from all over the World.

I am in the Carolinas for instance.

Of course if your profiles are not accurate then let please update them or let us know where you are located and someone will most likely come forward near you.

Bill.

[dabney121]

Thanks very wise advise and I agree we need help and training. I will look into it.

Can one of you take a stab at one more question?

What we are looking for this tiered disk assigning to resemble is:

Disk01 logical would maintain 2 physical disk addresses. For example they would be:
$1$DGA001:[directory]
$1$DGA002:wherever the new location isâ ¦we could name it $1$DGA002:[NRCTANDEM] so both naming conventions match

[Bill Pedersen]

The directory name on disk2 can be the same as on disk1. They do not intersect in any fashion.

As a way to get some BASIC OpenVMS training let me direct you to a free, on-line, self-paced environment:

http://plato.ccsscorp.com

I fully admit this is my own site but it is free for your use as well as for anyone else.

There are currently four courses which give you an introductory level experience on OpenVMS. It discusses command procedures, logicals, symbols, editors, file systems and more. There is also an on-line OpenVMS lab environment so you can experiment and test ideas without interfering with your own environments.

Good luck,

Bill.