Re: Disk Shuffle

On Apr 18, 3:42 pm, Bob Schlosser <rschlos...@xxxxxxxxxxxxxxxxxx>
wrote:
Wyell

The directory option will preserve your current accounting structure.



-----Original Message-----
From: HP-3000 Systems Discussion [mailto:HP300...@xxxxxxxxxxxxx]On

Behalf Of Wyell Grunwald
Sent: Friday, April 18, 2008 3:38 PM
To: HP300...@xxxxxxxxxxxxx
Subject: Re: [HP3000-L] Disk Shuffle

Matthew,

Thanks so much for your nuggets of wisdom !
The comments on the CONFIG group files being on LDEV=1 are great !
The primary path is not going to change (they are all SCSI connected, and I
am swapping the SCSI id's when I swap the drives).
One question, when I restore with the DIRECTORY option, does this make sure
that the access rights, capabilities, limits, etc. are the same as before ?
(That was why I was using BULDJOB1).

Wyell Grunwald

-------------- Original message --------------
From: Matthew Perdue <hp_3...@xxxxxxxxxxxxxxx>

To your store parameters I'd add:
PARTIALDB;ONVS=MPEXL_SYSTEM_VOLUME_SET
for your store tape set for the system volume set

PARTIALDB;ONVS=EXTRA_EXTERNAL_VOLUME_SET
for your store tape set for the external volume set

I'd make 2 store tape sets for each volume set and verify BOTH sets with
VSTORE

I'd change the config in SYSGEN in the group CONFIG to be what you want
*after* your move process, as the changes won't take affect until your
next
boot from the CONFIG group, which will be after you've moved discs. Unless

of course you boot between some of your steps, or issue a DOIONOW command
after you've changed the IO config, but you don't want to do that during
this move process anyway.

Just be sure *all* the files in your config group you intend to use are on

LDEV 1 after the install, otherwise you'll end up with a box that boots on

one drive (yep, LDEV 1) possibly without any other drives, and a bunch of
directory problems.

You don't need to purge all the accounts on the external volume set as the

VSCLOSE and SCRATCHVOL processes will do this for you (well, the
SCRATCHVOL
is actually the command that will do the work, the VSCLOSE makes the
SCRATCHVOL possible as you can't scratch an open volume set).

At #10 yes, change the SCSI id's.

Before #15 change the primary path in stable storage to reflect the new
primary path you want to use. This prevents boot problems in the future
when
it's possible to forget in haste that the primary path was changed when
the
discs were changed.

Step #26 not necessary.

On the restore for the system volume set and external volume set specify
the
DIRECTORY option, as well as the PROGRESS option set at 3 instead of 1 -
an
update every 3 minutes is usually sufficient.

When you do your restore and specify the DIRECTORY option you do not need
to
run the BULDJOB1 as the restore recreated the system directory for you.
You
should run the BULDJOB2 to take care of the udc's, unless you restored the

COMMAND.PUB.SYS file which is where the udc assignments are kept.

---original message---
Disk Shuffle - sounds like some sort of Adager command !
I have a HP 3000 928RX that has 6.5 on it with the following disk
configuration:
9 Gb internal drive - LDEV 1 (only 4 Gb recognized)
MPEXL_SYSTEM_VOLUME_SET
9 Gb internal drive - LDEV 2 (all 9 Gb recognized) MPEXL_SYSTEM_VOLUME_SET

1.9 Gb external drive - LDEV 523 EXTRA_EXTERNAL_VOLUME_SET
1.9 Gb external drive - LDEV 524 EXTRA_EXTERNAL_VOLUME_SET
I would like to go through the nasty process of switching the LDEV 1 drive

with the LDEV 523 drive, so that I have use of the 4 Gb of space that is
unrecognized (no, I cannot go to MPE 7.5). All drives are SCSI connected..
Does this sound like I could do this ? Below is a sketching of the steps
to
do this - did I leave anything out ?
1) Make sure that all the essential files can fit on 1.9 Gb (which I think

they will according the the OS installation manual)
2) Run BULDACCT @
3) Stop all jobs, background processes, network, etc.
4) STORE @.@.@;*T;DIRECTORY;SHOW to 2 separate tapes
5) Purge all ACCOUNTS on EXTRA_EXTERNAL_VOLUME_SET
6) VSCLOSE EXTRA_EXTERNAL_VOLUME_SET
7) VOLUTIL, SCRATCHVOL both disks in EXTRA_EXTERNAL_VOLUME_SET
8) SYSGEN, IO, DDEV 523, DDEV 524, HO, E, KE, TAPE (make 2 SLT tapes), E
9) ctrl-A SHUTDOWN, power off
10) physically swap the LDEV 523 disk with the LDEV 1 disk
11) swap the SCSI addresses on the disks in step 10 above ?
12) disconnect SCSI cable for external drives
13) boot the box from SLT tape created above in step 8
14) INSTALL
15) Reboot box from new LDEV 1 (primary path)
16) ISL> START NORECOVERY NOSYSSTART CONFIG=CONFG9X8 (This is so that it
will not remember the old disk configuration)
17) ctrl-A, SHUTDOWN
18) re-attach SCSI cable for external drives
19) Reboot box from primary path
20) ISL> START NORECOVERY NOSYSTART
21) SYSGEN, IO, add all paths, devices back in, HO, E, KE, E
22) ctrl-A, SHUTDOWN
23) Reboot box from primary path
24) ISL> START NORECOVERY NOSYSSTART
25) VOLUTIL, NEWSET command to re-create EXTRA_EXTERNAL_VOLUME_SET
26) SHUTDOWN / Reboot neccessary ?
27) From store tape created in step 4 above, RESTORE
*T;BULD@.PUB.SYS;SHOW;OLDDATE
28) OPENQ LP, STREAMS 10, Stream BULDJOB1.PUB (To re-create all accounts,
users, groups)
29) From store tape created in step 4 above, RESTORE
*T;@.@.@;SHOW;OLDDATE;KEEP;PROGRESS=1;CREATE
30) STREAM BULDJOB2.PUB (To re-create all udc's)
31) SHUTDOWN / Reboot WITH SYSSTART
What a monumental task. Please comment if I left something out !
Wyell Grunwald

* To join/leave the list, search archives, change list settings, *
* etc., please visithttp://raven.utc.edu/archives/hp3000-l.html*

* To join/leave the list, search archives, change list settings, *
* etc., please visithttp://raven.utc.edu/archives/hp3000-l.html*

https://www.mailcontrol.com/sr/wQw0zmjPoHdJTZGyOCrrhg==
slsDRt+7BL!7qk8DuIzVBW8xr7FDnd0701Q6O9SzTb9Cm8w9rID4pHOM8uEZ6cYqdoUyXaJMBQv­i
ocSk!!v+vAFTBddCvT2dqsnhCHwhpRIxFJrIvH0HJapznixvO6umAIiBln!rOs0IJHcWwCSABBQ­d
I!WV3Xo5YN!S8X2ZFxrVhykbxl2  
Click on the link above to report this email as spam to BlackSpider.

***************************Disclaimer***************************
The contents of this e-mail and any file transmitted with it are confidential and intended solely for the individual or entity to whom they are addressed.  The content may also contain legal, professional or other privileged information. If you received this e-mail in error, please destroy it immediately.  You should not copy or use it for any purpose nor disclose its contents to any other person. The views stated herein do not necessarily represent the view of the Company. Please ensure you have adequate virus protection before you open or detach any documents from this transmission. The Company does not accept any liability for  viruses.
Premier Farnell plc
150 Armley Road Leeds
LS12 2QQ
Telephone +44 (0) 870 129 8608
Fax +44 (0) 870 129 8610
Registered in England
Company Number 876412
Registered Office: Farnell House, Forge Lane, Leeds LS12 2NE ************************************************************

* To join/leave the list, search archives, change list settings, *
* etc., please visithttp://raven.utc.edu/archives/hp3000-l.html*- Hide quoted text -

- Show quoted text -

Wyell,

A couple of things. During a recent recovery, when I lost my system
disc, I ran:

:RESTORE *T;;DIRECTORY;CREATE=CREATOR;SHOW=OFFLINE

and I KNOW my full backup has 'directory' in it, and it DID NOT
restore my accounting structure correctly. (yes, I have the updated
version, I'm on 6.5 as well). Everything was restored, but the group
and account security was not restored correctly. I would run BULDACCT
and create the BULDJOB's anyway just in case.

Second, when you do run BULDACCT, check BULDJOB1 and make sure all of
your accounts are there. It builds it alphabetically, so just do a
search on "NEWACCT" and make sure your last account is there. I run my
BULDACCT in my full backups and build the output files first. However,
if you don't build the files big enough, BULDACCT does not give any
kind of error, it just stops!
I built my BULDJOB1 file at 200,000 records and it wasn't large
enough. It built the job stream up to record 200000, which was about
2/3 of what I needed (it only got to the M's) and then just stopped,
no EOF warning or error, nothing, and then continued on with the
backup.

Good luck.

Dan
.

Relevant Pages

  • Re: [PATCH 1/3] accounting: task counters for disk/network
    ... and process level counters for CPU, memory, disk and network utilization. ... The current kernel offers the system level counters, ... The kernel feature "task I/O accounting" currently maintains ... I/O request is accounted to one of the processes which can be very useful ...
    (Linux-Kernel)
  • Re: [Patch][RFC] Disabling per-tgid stats on task exit in taskstats
    ... - it repeatedly does recvto retrieve data from kernel ... disk as it would be for an accounting daemon. ... BSD accounting and the CSA writes accounting data from ... doing the recv() and only write to disk the data it wants? ...
    (Linux-Kernel)
  • Lots of accounting data
    ... I'm playing a game of cat and mouse with process accounting and disk ... I built some boxes with 9GB /var partitions, ...
    (freebsd-questions)
  • Re: I/O statistics per process
    ... chars-written accounting. ... disk - especially if there is more than one task consuming cpu. ... The per-process chars-read and chars-writeen accounting is made ... The patch provides accounting data only at ...
    (Linux-Kernel)
  • Re: Secondary IDE slave
    ... >> with one disk configured as slave.... ... >> config) to make use of CSEL, then it results into a single disk ... >> as slave. ... >>> the CompactFlash connector is hardwired to the SECONDARY IDE ...
    (microsoft.public.windowsce.platbuilder)