"Antognini, David" wrote:
> 
> The latest Backup exec has the proper client for Linux. Its vital to have a
> backup that is NOT on a server. If someone breaks in and all you have are a
> bunch of tar files on various servers then they could totally wipe you out.
> If you have it on tape then at least you could recover.

Thats why you back up to tape and physically remove the tapes from the
area/building containing the computer(s), i.e. a safe deposit box, fire
proof safe, etc...
 
> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, July 08, 1999 4:20 PM
> To: [EMAIL PROTECTED]
> Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED];
> [EMAIL PROTECTED]
> Subject: Re: backup/redundancy solutions
> 
> hi allan, steve,
> 
> using NT to backup linux destroys linux permissions...
> using linux to backup NTs destroys NT permissions/uid/gid
> 
> - as others have stated earlier
> using mirror'ed disks might result in the original disk being bad/corrupted
> and the mirror itself to be bad...( you have the time between mirrors
> to find any corruption or bad data etc...
>         - I keep a tar file ( unmirrored ) so that at least one day of
> backup
>         file will/should have a last known good file
> 
> - problem with 1 huge backup machine is if that one machine dies...
>   your raid5 backup is dead to ?? ( bad power surge )
>         - that huge disk/system sits idle 99% of the time ??

The Raid5 backup might die, but you still have a tape backup of the
raid5 which is at most 1 day old, and the original machines still
contain the original data. The "huge backup machine" also can perform
other server tasks such as DNS & NIS... Besides, as you mentioned below,
20GB ide drives are relatively cheap, so having 100GB worth of them
*only* being used as a "backup storage" seems like pretty cheap
insurance...compared to the value of the information preserved - what
does it cost to "lose" one week of work by 5 programmers?

> 
> - tape backups for offsite archives only... too slow to recover files
>         - 20Gb ide drive is $300...( fast, cheap, good for online mirror,
> etc )

So you use a combination of huge disk for immediate backup, then tape to
back up the backup disk... as you mentioned big IDE drives are cheap,
get a bunch of them for the backup area...

> 
> - we have 3 - 64Gb hardware raid as /home... ( or equivalant )
>         and the backups of it spread across 4 the other servers
> 

We do just the opposite...we have /home (or equivalent) spread across 10
machines and a 5x24Gb IDE software raid as backup. We felt having the
user's "home" on their machine would provide better performance since
most of their work is done on their own machine...

> - I keep a tar file of changes...
>         - the tar file is used to create the mirror periodically
>                 - hourly backups to keep a copy of hourly changes
>                 - daily backups/mirror to mirror todays work
>         - the tar file will have the "changes" for however long we keep the
> data
>                 - if a file changes daily....going back in time, we should
>                 find a uncorrupted file...but...not to to date ?
> 

Sounds like a good strategy!

> - backup can also be corrupt or incomplete
>         - usually NFS problems
>         - disk 100% full problems
>         - lack of backup permission to remotely read root protected files
> 
> - if the full backup or incremental backup is bad... any subsquent backups
> are
>   bad too ...
>         - my daily incremental backups start for the last full backup
>         - my weekly backups span 30 days and a  full backup
>                 ( if this week full backup is bad...30-day incremental can
>                 (  rebuild/recover from last weeks full backup
>         - weekly full backups reside on different disks on 3 different
> machines
>         - it auto-rotates to different backup machines...
>                 ( backup_1, backup_7, backup_30, backup_wk1, backup_wk2,
> etc..
> 
> - i wrote my own backup script to do backups and "mirror"
>         - have search capability too
> 

Sounds like your approach is a good one...we basically use disk for the
"hourly" backups, and tape for the daily/weekly backups...


> - some people use cdr to back really critical data
> 

Yup...do that too, particularly for the code we write and the individual
project deliverables...

John

-- 
John Burton, Ph.D.
Senior Associate                 GATS, Inc.  
[EMAIL PROTECTED]          11864 Canon Blvd - Suite 101
[EMAIL PROTECTED] (personal)          Newport News, VA 23606
(757) 873-5920 (voice)           (757) 873-5920 (fax)

-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [EMAIL PROTECTED]

Reply via email to