[JA] Freelists is back
- From: jim.henderson@xxxxxxxx
- To: juno_accmail@xxxxxxxxxxxxx
- Date: Thu, 12 Jul 2001 10:50:20 -0400
>JH >No hard drives in cold storage?
BLJ> Well, what happened there was that Avenir (the computer
> running Freelists) runs RAID arrays, where data is mirrored...
Sure. All drives in the array are hot, and are thus subject to viruses,
lightning strikes and other hazards. That's why there is cold storage.
Keep a set of obsolescent hardware and backup tape/disc/whatever on a
shelf in another building, to rebuild in case flood, burglar or other
hazard takes out the working hardware.
BLJ> If you'd like to see a blow-by-blow, you can read
> //www.freelists.org/log.html
Thanks, Brian; I'll take a look.
> Subject: [JA] Re: Freelists is back
> From: Billy Wong <bhs1989@xxxxxxxx>
>JH > Lightning killed my modem last summer.
BW> And is this with or without a surge protector?
I had two USR v.everything modems working that morning. The one doing
Web duty had its phone line on a Isotel surge suppressor from TrippLite;
the one doing BBS duty had its power line protected but not its phone
line. The lightning strike nearly knocked me out of my chair. More
important, it killed the surge suppressor and the unprotected modem.
BW> I bought a new SP that has telephone line
> protection included and have kept the SP switched on at all times.
No need to keep it turned on. Surge suppression is entirely passive;
it's always on duty regardless of the position of the power switch.
On the other hand, many surge suppressors are ineffective, regardless.
Perhaps the majority don't actually work. Isotel from Tripplite is one
brand that does the job. It costs about thrice the price of a suppressor
that fails to suppress, but I figure anything ineffective is overpriced.
I no longer use an expensive external modem for Internet; that job is
done by a no-name internal modem, which incidentally is faster. One of
my surviving expensive modems is in cold storage in case the BBS needs
it.
Since a bunch of people depend on my providing reliable service, I use
multiple methods to assure it: 1) A second hard drive inside the BBS
machine, with a BATch file running automatically every morning to backup
the most vital files that make the BBS work. 2) A hot standby machine on
the same desk, with a working copy of the BBS, updated every few days.
3) Software and file backups taken every few weeks, stored off site, and
tested four times a year. 4) Hardware, suitable for running the BBS with
limited service, on a shelf in a cool dark place far from electrical
outlets.
Method 1 has sometimes failed for me, but I won't give it up since it
sometimes works and costs little effort. Method 2 hasn't yet failed me.
Some of my Method 3 backups have failed, but others worked. Method 4
once failed me but it was all right as I could go to the computer store,
buy replacement hardware, restore my backup and return to business that
day. Someone who uses hardware not stocked at the neighborhood computer
store should make sure of the spares in their own stock, rather than take
the risk of having to wait for a delivery service.
RAID 0 redundancy provides instant protection when one hard drive dies.
That's a good thing, but too expensive for me. People who pay for that
kind of protection should bear in mind that a single dead drive is not
the only or even the biggest danger, which means RAID only does a partial
job. Multiple hazards call for multi- layered protection.
To unsubscribe, send a message to listar@xxxxxxxxxxxxxxxxx with
"unsubscribe juno_accmail" in the body or subject.
OR visit http://freelists.dhs.org
~*~
Other related posts:
