ADSM-L

Re: Disaster-Recovery

2000-05-10 15:36:33
Subject: Re: Disaster-Recovery
From: "Cook, Dwight E" <cookde AT BP DOT COM>
Date: Wed, 10 May 2000 14:36:33 -0500
Now I do have some pools that the clients may use which I do perform  a
backup on that storage pool...
They have to specifically bind their data to the management class that
points there and I only allow archives to go there.
All the tapes remain inside the ATL
and I've even (for a short period of time) made available storage way away
on a remote server so the initial backups were in a different site within
the same state and the backup of the storage pool was on a different server
in another state... but that OH MY GOD CRITICAL APPLICATION was deemed
useless and the server went away and so did my jumping through hoops to back
it up
        ;-)
This is what I really like about ADSM/TSM... every time I've been asked if
it would be possible to do BLAH... I've been able to perform BLAH with
adsm/tsm

My way is just A way...application owners request their way, I tell them how
much their way costs (tapes, network, processor), they say NO WAY, and we
figure out a HALF WAY that makes everyone happy... but at the base of it all
is the same 'ol adsm/tsm.

Dwight

> ----------
> From:         Lawrence Clark[SMTP:Larry_Clark AT THRUWAY.STATE.NY DOT US]
> Reply To:     ADSM: Dist Stor Manager
> Sent:         Wednesday, May 10, 2000 2:15 PM
> To:   ADSM-L AT VM.MARIST DOT EDU
> Subject:      Re: Disaster-Recovery
>
> Dwight:
>
> I was wondering what you would do if critical data on your application
> server was destroyed and you found out that you have a media error on the
> only copy of your backup........
>
> >>> cookde AT BP DOT COM 05/10/00 02:48PM >>>
> I take daily full adsm data base backups... so if the backup from the
> previous day is bad, I shift back a day... and so on...
> BUT I need to point out that if I have to go back more than a couple of
> days
> the amount of audit volumes I'd need to run to get things back in sync
> would
> basically invalidate my efforts... that is I can't allow the backup
> environment to be down that amount of time.
> These are strictly CYA backups for all the clients to protect against
> finger
> check deletions, data corruption, a disaster in the main computing
> facility
> and the such in their environments.
> If my adsm environment totally blows it would be more important to get a
> blank environment up and running and get a fresh backup from all the
> clients
> than to recover the lost info within ADSM.
> Not necessarily the wrong or right way to do things... just the way we do
> things...
>
> Dwight
>
> > ----------
> > From:         Lawrence Clark[SMTP:Larry_Clark AT THRUWAY.STATE.NY DOT US]
> > Reply To:     ADSM: Dist Stor Manager
> > Sent:         Wednesday, May 10, 2000 12:30 PM
> > To:   ADSM-L AT VM.MARIST DOT EDU
> > Subject:      Re: Disaster-Recovery
> >
> > But what if your server fries and you try to recreate it from cartridges
> > with media errors? Isn't that the point of (offsite) copypool data?
> >
> > Larry clark
> > NYS Thruway Authority
> >
> > >>> cookde AT BP DOT COM 05/10/00 08:51AM >>>
> > No I don't use DRM...
> > Environments deemed business critical backup to adsm environments
> offsite,
> > since the backups are already offsite we don't perform backups of our
> > backups.
> > I do gather the information required to put the adsm server back
> together
> > in
> > the event of a burnout of the adsm server as long as the tapes in the
> atl
> > still exist. If my tapes melt that is the end of it all... except that
> the
> > applications that are backed up there are still running and I point them
> > to
> > another environment until I get the lost one rebuilt/replaced.
> >
> > Dwight
> >
> > > ----------
> > > From:         Mr. Lindsay Morris[SMTP:lmorris AT OPENMIC DOT COM]
> > > Reply To:     ADSM: Dist Stor Manager
> > > Sent:         Monday, May 08, 2000 11:19 PM
> > > To:   ADSM-L AT VM.MARIST DOT EDU
> > > Subject:      Re: Disaster-Recovery
> > >
> > > You had these problems using DRM?  Or not?
> > >
> > > "Cook, Dwight E" wrote:
> > >
> > > > Now I haven't tried it lately (busy as a bee already) BUT
> > > > the last time I was working on bringing up a recovery environment of
> > my
> > > adsm
> > > > server these were my thoughts and experiences...
> > > > OK, I don't need my big DISKPOOL's until all the clients that were
> > wiped
> > > out
> > > > also, get restored...
> > > > Soooo if I can just bring up an adsm environment with the DB, LOG,
> > > LIBRARY
> > > > and all the LIBVOLS I should be good to go to restore clients...
> > > > WRONG !! ! ! ! ! !
> > > > I would go to start ADSM and it kept trying to bring on storage
> > volumes
> > > that
> > > > existed in the production environment
> > > > and after some limit of errors was reached ('cause the diskpool
> > volumes
> > > were
> > > > not around) it aborted starting the server !
> > > > OUCH !
> > > >
> > > > this is why I say get another machine and just try to restore your
> > adsm
> > > > environment... it is a LARGE learning experience...
> > > >
> > > > Dwight
> > > >
> > > > (and no you really don't need identical machines... but why make
> life
> > > harder
> > > > than it has to be)
> > > >
> > > > > ----------
> > > > > From:         Lawrence Clark[SMTP:Larry_Clark AT THRUWAY.STATE.NY DOT 
> > > > > US]
>
> > > > > Reply To:     ADSM: Dist Stor Manager
> > > > > Sent:         Tuesday, May 09, 2000 1:59 PM
> > > > > To:   ADSM-L AT VM.MARIST DOT EDU
> > > > > Subject:      Re: Disaster-Recovery
> > > > >
> > > > > I would think once your mksysb is restored you would have
> everything
> > > but
> > > > > the database, assuming its all on rootvg.
> > > > >
> > > > > But is it true you do not need an identical machine?
> > > > >
> > > > > I'm wondering:
> > > > >
> > > > > If I have a ADSM server with both disk and cartridge based storage
> > > pools a
> > > > > tape library  for storage pools and a second tape library at a
> > remote
> > > site
> > > > > for the copy pool, am I still set up for a rabid recovery?
> > > > >
> > > > > Wouldn't I need to?:
> > > > > - have a server with similar disk allocation for the disk-based
> > > storage
> > > > > pools
> > > > > - have to remove all the copypool tapes from the second library,
> > > attach it
> > > > > to a box at the remote site and define it as the tape library that
> > was
> > > at
> > > > > the primary site
> > > > > - then mark the  volumes in the storage pools in the primary
> library
> > > as
> > > > > destroyed. But then wouldn't ADSM look for the copypool volumes in
> > the
> > > > > (now missing) tape library?
> > > > >
> > > > > >>> bill.jolley-eds AT EDS DOT COM 05/09/00 02:50PM >>>
> > > > > Once the mksysb, adsm db, devconfig,and volhist are restored,
> > > should'nt my
> > > > > server code be
> > > > > loaded?
> > > > >
> > > > > -----Original Message-----
> > > > > From: Lynch, Rich [mailto:Lynch.Rich AT MBCO DOT COM]
> > > > > Sent: Tuesday, May 09, 2000 2:23 PM
> > > > > To: ADSM-L AT VM.MARIST DOT EDU
> > > > > Subject: Re: Disaster-Recovery
> > > > >
> > > > >
> > > > > you really dont need a Identical machine, you can use cloner
> (mksysb
> > > and a
> > > > > cdrom) to boot up the recovery node and restore the node, even if
> > its
> > > > > totally different hardware. Weve done this at Sunguard and IBM DR
> > site
> > > a
> > > > > couple of times, using ADSM and 3590 drives to recover the rest of
> > our
> > > > > data
> > > > > (SAP)
> > > > >
> > > > > -----Original Message-----
> > > > > From: Jolley, Bill (Matlen Silver)
> [mailto:bill.jolley-eds AT EDS DOT COM]
> > > > > Sent: Tuesday, May 09, 2000 1:12 PM
> > > > > To: ADSM-L AT VM.MARIST DOT EDU
> > > > > Subject: Re: Disaster-Recovery
> > > > >
> > > > >
> > > > > Thanks, I been down this road before and no dice on identical
> > machine.
> > > > >
> > > > > -----Original Message-----
> > > > > From: Cook, Dwight E [mailto:cookde AT BP DOT COM]
> > > > > Sent: Tuesday, May 09, 2000 12:56 PM
> > > > > To: ADSM-L AT VM.MARIST DOT EDU
> > > > > Subject: Re: Disaster-Recovery
> > > > >
> > > > >
> > > > > Ok, I can't hand out documentation... (some of that lawyer stuff)
> > > > > What I'd do (actually what I did)
> > > > > tell them you need a test/recovery server identical to your
> existing
> > > > > machine
> > > > > perform a recovery... OH WHAT AN EYE OPENER IT WILL BE !
> > > > > there will be so much stuff you'll need that you never even
> thought
> > of
> > > !
> > > > > then just make sure and set up jobs in your production
> environments
> > > that
> > > > > will gather the info you've discovered you need to recover the
> > > > > environment...
> > > > > it will be ugly the first time but I thought it was rather fun...
> > (ok,
> > > I'm
> > > > > twisted)
> > > > >
> > > > > later,
> > > > >         Dwight
> > > > >
> > > > > > ----------
> > > > > > From:         Jolley, Bill (Matlen
> > > Silver)[SMTP:bill.jolley-eds AT EDS DOT COM]
> > > > >
> > > > > > Reply To:     ADSM: Dist Stor Manager
> > > > > > Sent:         Tuesday, May 09, 2000 10:09 AM
> > > > > > To:   ADSM-L AT VM.MARIST DOT EDU
> > > > > > Subject:      Disaster-Recovery
> > > > > >
> > > > > > I have been asked for a procedure to recovery my adsm server
> > > 3.1.2.40 os
> > > > > > level 4.3.1 (S70) atl 3494 3590 tape drives.
> > > > > > Do anyone have a dr plan in place or where can I get dr info?
> > > > > >
> > > > > > Thanks
> > > > > >
> > > > >
> > >
> > > --
> > > Mr. Lindsay Morris
> > > Certified: AIX,ADSM, TSM, HACMP,SP
> > > Gresham Enterprise Storage
> > > lmorris AT openmic DOT com
> > > 606-253-8000
> > >
> >
>
<Prev in Thread] Current Thread [Next in Thread>