CS 200 Spring 2019
1
Backups
CS 200
System Management
Backups
06 May 2011
Let’s discuss the facts of life…
You will die. Your disk will die.
Your disk will die first.
CS 200 Spring 2019
2
Backups
Who does them?
The user, manually
at day’s end, make copies on …
and, if you’re smart, before you make major changes
The user, by running a program, based on time-of-last-modification eg Retrospect, Norton Fastback, SmartBackup
Your PC, or a network server, by automatically running an application, based on time-of-last- modification
eg Retrospect + Retrospect Client
CS 200 Spring 2019 3
Backups
Terminology (1)
A Full Backup
Back up everything, whether it’s changed or not
Incremental
Back up only files changed or created since the last backup
Duplicate sets
Make n copies of each backup (n > 1)
Rotating sets
Rotate through n sets of backup tapes/CDs/DVDs/… (n > 1)
CS 200 Spring 2019 4
Backups
Terminology (2)
Disk mirroring
Replicating the contents (or changed contents, & deleting deleted files) of one disk to another
Incremental mirroring
Copy changes to a duplicate disk
Offsite storage
Far enough away as to eliminate the risk of one event destroying both your PC and the backup
Archival storage
Periodically take a backup set out of service & “archive it”
CS 200 Spring 2019 5
Backups
A Weekly-Cycle Incremental Backup Strategy
Monday
Tuesday
Wednesday
Thursday
Friday
Full
Incremental 1
Incremental 2
Incremental 3
Incremental 4
Full
Incremental 1
Incremental 2
Incremental 3
Incremental 4
Full
Incremental 1
Incremental 2
Incremental 3
Incremental 4
Full
Incremental 1
Incremental 2
Incremental 3
Incremental 4
•••
•••
•••
•••
•••
To recover an individual file
work backwards (4, 3, 2, 1, Full) until you find it
To restore the entire file system
work forwards (Full, 1, 2, 3, 4) restoring everything
CS 200 Spring 2019 6
Backups
More Terminology
If you keep a “catalog” of what’s in each backup
ie a table-of-contents (of file paths, size, dates of creation/modification, location in backup)
you could search that
to identify and restore the most recent version of each file
instead of copying them all & over-writing older versions
and you could likely keep the catalog on disk, making retrieval of an individual file much faster A “session” is a particular backup (eg the Week 2, Weds incremental backup)
A “snapshot” is exactly what’s on your disk at the time of a backup
Files not changed since the last previous backup are in the snapshot
but not in the session
CS 200 Spring 2019 7
Backups
A Duplicate Set (2) Incremental Backup Strategy
Monday
Tuesday
Wednesday
Thursday
Friday
Full A Full B
Incremental
A1
Incremental
A2
Incremental
A3
Incremental
A4
Full A Full B
Incremental Incremental
A1
Incremental Incremental
A2
Incremental Incremental
A3
Incremental Incremental
A4
Full A Full B
Incremental Incremental
A1
Incremental Incremental
A2
Incremental Incremental
A3
Incremental Incremental
A4
Full A Full B
Incremental Incremental
A1
Incremental Incremental
A2
Incremental Incremental
A3
Incremental Incremental
A4
•••
Incremental •••
Incremental •••
Incremental •••
Incremental •••
If one backup goes bad, you have another equally recent backup
If lightning fries your machine AND the backup you are making,
you can restore from the other backup, and it’s equally recent
But … it takes twice as many tapes / disk(ette)s, and twice the time
CS 200 Spring 2019 8
Backups
A Rotating Set (2) Incremental Backup Strategy
Monday
Tuesday
Wednesday
Thursday
Friday
Full A
Incremental A1
Incremental A2
Incremental A3
Incremental A4
Full B
Incremental B1
Incremental B2
Incremental B3
Incremental B4
Full A
Incremental A1
Incremental A2
Incremental A3
Incremental A4
Full B
Incremental B1
Incremental B2
Incremental B3
Incremental B4
•••
•••
•••
•••
•••
If the most recent backup goes bad, you have another,
though you will lose recently created files and recent changes
If lightning fries your machine AND the backup you are making,
you can restore from the other backup
An advantage — if you’re infected by a virus
and some infected files were backed up
you can get older, uninfected versions from the older backup
No more time / work is involved
but … it takes twice as many tapes / CDs / DVDs / disks / …
CS 200 Spring 2019 9
Backups
A Duplicate Set (2) + Rotating (2) Incremental Backup Strategy
Monday
Tuesday
Wednesday
Thursday
Friday
Full A Full B
Incremental
A1
Incremental
A2
Incremental
A3
Incremental
A4
Full C Full D
Incremental Incremental
C1
Incremental Incremental
C2
Incremental Incremental
C3
Incremental Incremental
C4
Full A Full B
Incremental Incremental
A1
Incremental Incremental
A2
Incremental Incremental
A3
Incremental Incremental
A4
Full A Full B
Incremental Incremental
C1
Incremental Incremental
C2
Incremental Incremental
C3
Incremental Incremental
C4
•••
Incremental •••
Incremental •••
Incremental •••
Incremental •••
Combines the advantages of both
But takes more time, and more tapes / CDs / DVDs / …
CS 200 Spring 2019 10
Backups
Data Compression
x 2 (on average) is easy
many backup programs use data compression
must use the backup program to retrieve a file
whereas with a file-system-formatted backup
you just mount the disk & use standard file copying techniques
A related issue: O/S-format or proprietary format
disks are allocated in 512-byte “allocation blocks” (for efficiency)
some backup programs
create one gigantic file &
write your files inside it,
keeping track of file locations themselves
so as not to waste disk space
but then must use the backup program to retrieve the file;
if a piece of your disk / CD / DVD goes bad, or the vendor out of business, you’re probably up the creek
CS 200 Spring 2019 11
Backups
Security
Do you want a backup program (or file archiving program) that offers (optionally) encrypted storage?
(optionally) encrypted data transmission (if done over a network)? how good should the encryption be?
CS 200 Spring 2019 12
Backups
File > Save… Backups
Suppose your application crashes while saving your work?
When you select File > save, many apps will optionally
change the name of the existing document file to something like “Backup of …”,
then write your current document to a new file on disk
MS Word, for example:
Backups
CS 200 Spring 2019 13
Preferences: Save
Ditto For Excel
CS 200 Spring 2019 14
Backups
And so on…
CS 200 Spring 2019 15
Backups
And for TextWrangler
⌘ – S (Mac) / Ctrl– S (Windows)
Train yourself to do a File > Save
automatically
every time you pause to think between changes
CS 200 Spring 2019 16
Backups
Folder Archives
When you’re working with multiple files in a folder
Periodically make a compressed archive of the folder’s contents
eg in Mac OS X’s Finder, select the folder, then select File > Compress “• • •” eg in Windows XP’s Explorer, File > Send To > Compressed (zipped) Folder or the analogous items on the contextual (right-click) menus in each system
CS 200 Spring 2019 17
Backups
Numbering These “Checkpoints”
Numbered files (after a File > Save)
someDocument_1.doc
someDocument_2.doc
someDocument_3.doc
•••
Numbered archives
someFolder_1.zip
someFolder_2.zip
someFolder_3.zip
•••
Sometimes it’s useful to keep a list of what you’d just finished in each checkpoint
CS 200 Spring 2019 18
Backups
Time Machine (OS X 10.5 & Later)
Uses a separate disk (or partition)
Keeps
hourly backups for the past 24 hours
daily backups for the past month
weekly backups until your backup disk is full
Each backup looks like an exact copy of your disks;
actually, “hard links” are used so there’s only one copy of each version
Effectively a full backup + incrementals
Multiple (simultaneous) sets not possible
You can (manually) change backup disks
⇒ rotating backups and off-site archives are possible
CS 200 Spring 2019 19
Backups