HP3000-L Archives

March 1998, Week 3

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Mark Bixby <[log in to unmask]>
Reply To:
Date:
Mon, 16 Mar 1998 11:52:22 -0800
Content-Type:
text/plain
Parts/Attachments:
text/plain (112 lines)
Based on my loooong day yesterday doing my defrag re-install, here is an
updated checklist the for procedure I'll be using in the future:

                          MPE/iX Re-Install Procedure

___  :XEQ SHUTDOWN.PUB.CCCD from OPERATOR.SYS

___  :STREAM BACKSLT.OPERATOR.SYS (twice if paranoid)

___  :XEQ CHECKSLT.MPEXL.TELESUP for each SLT backup created above.

___  :STREAM BACKUP.INSTALL.CCCD (twice if paranoid)

Which does in part:

!  CONTINUE
!  DEVCTRL.MPEXL.TELESUP 7,DISABLE,ENABLE
!  CONTINUE
!  DEVCTRL.MPEXL.TELESUP 8,DISABLE,ENABLE
!  CONTINUE
!  STORE ^INDSTORE;;STORESET=(*FULLSET1),(*FULLSET2);&
!    ONVS=MPEXL_SYSTEM_VOLUME_SET,CCCD_MIRROR,CCCD_JUNK;&
!    DIRECTORY;INTER;SHOW=LONG;PROGRESS=5;MAXTAPEBUF;ONLINE;STATISTICS;&
!    COMPRESS=HIGH

My normal system backups are online backups (concurrent batch activity occurs)
done with DDS-2 hardware compression only (i.e. no COMPRESS=HIGH).

I did some benchmarks prior to the re-install day, and discovered that
COMPRESS=HIGH *without* hardware compression results in a substantial 33%
backup runtime decrease.  This improvement comes at the expense of a 30-50%
CPU load on my 969KS200, so I don't think I'll switch all of my full backups
to this method because it would slow batch production.

INDSTORE is ordered so that important files appear at the front of the backup
so they may be restored quickly:

@.INSTALL.CCCD
@.UDC.CCCD
COMMAND.PUB.SYS
EDITOR.PUB.SYS
@.PUB.SYS-COMMAND.PUB.SYS-EDITOR.PUB.SYS
@[log in to unmask]@.PUB.SYS
@[log in to unmask]@[log in to unmask]@[log in to unmask]@.SYS

___  :STREAM VERIFY.INSTALL.CCCD for each full backup created above.

Which does in part:

!VSTORE ;@[log in to unmask]@;SHOW=LONG;ONERR=QUIT;RESTORESET=(*SET1),(*SET2);&
! NODECOMPRESS;PROGRESS=5

I've never had logical corruption on a store tape, so there's no need to
decompress the data.  I'm just interested in media readability.

___  =SHUTDOWN

___  ISL INSTALL from the SLT tape created by BACKSLT.

___  :HELLO MANAGER.SYS,PUB

___  :VSCLOSE CCCD_MIRROR and CCCD_JUNK non-system volume sets

___  :RESTORE *T;CCCDVOLS.INSTALL.CCCD,@.PUB.SYS;SHOW;LOCAL;KEEP;OLDDATE;DEV=1

___  :XEQ VOLUTIL.PUB.SYS "USE CCCDVOLS" to recreate volume set configuration.

CCCDVOLS was created by my BULDVOL script; I manually inserted a bunch of
scratchvol commands.  BULDVOL successfully determined my full configuration,
which was recreated successfully by VOLUTIL.

___  :RESTORE *T;;DIRECTORY;SHOW;OLDDATE

Here something weird happened:

/SYS/PUB/public_html      COULD NOT BE RESTRICTED TO THE SYSTEM MASTER VOLUME
/SYS/hpstage              ditto
/SYS/hpstage/MPEKX56A     ditto
/SYS/hpstage/base-archive ditto
/SYS/hpstage/export       ditto
/SYS/hpstage/import       ditto

Not a big deal; I ended up purging public_html, and the stage stuff can be
recreated via STAGEMAN.

But still weird.  Does anybody have an explanation?

___  :RESTORE *T;@.INSTALL.CCCD,@.UDC.CCCD;SHOW;OLDDATE

___  :STREAMS 10

___  :OPENQ LP

___  :STREAM RESTJOB.INSTALL.CCCD

Which does in part:

!RESTORE ;@[log in to unmask]@;RESTORESET=(*SET1),(*SET2);&
! SHOW=LONG;CREATE;KEEP;OLDDATE;ONERR=SKIP;PROGRESS=5

___  Manually deal with files that had creator problems.

___  =SHUTDOWN

___  ISL START NORECOVERY
--
Mark Bixby                      E-mail: [log in to unmask]
Coast Community College Dist.   Web: http://www.cccd.edu/~markb/
District Information Services   1370 Adams Ave, Costa Mesa, CA, USA 92626-5429
Technical Support               Voice: +1 714 438-4647
"You can tune a file system, but you can't tune a fish." - tunefs(1M)

ATOM RSS1 RSS2