News Group Posts

removable backup

From: brian@aljex.com (Brian K. White)
Subject: Re: backup/restore using disk instead of tape
Date: 18 Nov 2002 20:17:09 -0800
References: <60bd4c6b.0211161339.69262ce2@posting.google.com>
<60bd4c6b.0211171317.4abbfc19@posting.google.com>
<H5qvw8.v1o@wjv.com>
<60bd4c6b.0211180134.5801a864@posting.google.com>
<H5s864.18rK@wjv.com> 

bv@wjv.comREMOVE (Bill Vermillion) wrote in message

> It needs more than padding.  Cameras tend to be more rugged than HD
> devices.  Be sure to check the specs on the HD drives you use
> for static G loads.  If they really want to do this, then
> the best thing would be to get the 2.5" drives designated for
> lap-tops as most of them will take a 100G shock  when turned off.
> That is a lot considering the human body will only survive 20Gs.

wow that is a *great* idea. do they make 150 gig scsi laptop drives?
I'd happlily make up sca2 adapters for them if they are even
available.
 

(Editors note): Of course one issue here is that ide drives aren't hot swappable- you'd have to bring the system down for this to work

ok that sounds facetious, but I actually do like the laptop drive idea
because I hadn't thought of it and it would help a lot with
portability and ruggedness.

actually, ... this may be an even better answer than using the scsi
hot swap bays. there are 3 unused ordinary 5 1/4 inch bays and the
motherboard IDE only being used by the cdrom. I could put the other
ide channel on a standard removable ide 5 1/4 inch dirve bay, and buy
extra bays to get the removable shuttle part to mount 2.5 inch drives
in. I could "mount" the 2.5 drives inside the 3.5 inch shuttles by
just packing them with foam on all sides, and the user could just
carry the bare shuttles without any extra padding necessary.

it's dirt cheap to replace the drive bay or a shuttle when the
connector starts to wear out too, unlike the sca2 bays where one side
of the connector is an integral part of a several hundred dollar
drive, and the other connector is part of a back plane that may not be
able to be replaced except by buying a whole new server case. (5u rack
mount w/ dual power and I don't even know how many fans...). Probably
the drive side connector will never wear out because that is not the
spring-loaded side, but still...

maybe I can play the "drives will get bigger" game and get 60 gig
drives now, banking on the hope that by the time the actual compressed
full backup reaches that size, new 2.5 inch drives will be available
at that time in higher capacities. Worse case scenario: they are not
and we switch to putting ordinary ide drives in the shuttles and
padding them externally for transport. which is still a little better
than what I'm contemplating now, because the drives would be cheap
enough that we can have lots more in live rotation, and the wear &
tear on the hot-swap connector is made unimportant by the ease with
which it can simply be replaced when it wears out.

I dislike the idea of using ide though. I don't know how much it would
impact the servers performance. I guess theoretically it shouldn't
matter since the backup should be running mostly by itself anyways.

> >and will have one or more extra drives besides the number that
> >will be in live rotation.
> 
> What will the total number of drives be.

unsure as yet. maybe just two at first. (that's all the other device
offered, as-shipped, though they would sell you more, and you had to
get them from them, because they needed to be specially
formatted...raising the cost even higher) we are currently rsyncing
the data to 2 other servers, one is 10 or 15 miles away. probably they
will get 5 for live rotation, and if one is damaged they will just use
the remaining 4 while the replacement is ordered. I don't see a lot of
necessity to have so many copies. For a tape it's a good idea because
tapes are kind of like floppies in that you shouldn't *really* depend
on the media being perfect, because it does suffer wear and loss
sometimes. the drives are fragile, but aside from dropping or zapping
with static, that data is basically 100% effective. (else servers
would be crashing and dying left and right on an hourly basis, and
they are not)

> 
> >We actually did not come to this idea in a vacuum. another $6000
> >"commercial backup solution" basically only offered this as the
> >answer for getting the backups off site.
> 
> That's pretty bizarre IMO.

yeah, I said as much to, slightly less politely :)
(yet after a while here I am attempting to do about the same thing,
just without the problematical whole extra server just to house the
drives)

> 
> > .... I'd done it using plain tar before in various odd
> >situations where it was the easiest way to move a lot of data
> >between two boxes or os's, and I knew ctar like plain tar would
> >try and write to any device you want.
> 
> Tape data is far more transportable than HD data.   As long as
> you are aware of that and will never need to have the data moved to
> another platform that works.

it would only ever concevably be open server, unixware, linux,
freebsd, or in a extremely abstract parallel univers, solaris. no
problems there.

there is no filesystem or partitions, simply a disk of (super)tar
data.
only windows would be baffled by it.


> And why not the AIT or VXA solutions. Those are cheaper but offer
> a lot of storage. VXA-1 comes in SCSI, Firewire and ATAPI. The
> VXA-2 is 80GB native/uncompressed. Ultra SCSI only an lists for
> $999.  It's basically a DATish concept but with packet writing to
> tape.
> 
> AIT-3 is 100GB native. That's the latest version. AIT-2 is 50GB.
> You mentioned compression and that many are. You will often find
> that hardware compression will outperform SW compression. The
> only way to be sure is test with your data to see if it offers
> enough gain.  Tapes are $66 each [Pricewatch price}. Drives are in
> the low $3K range.  It's a good target when DAT is too small
> and LTO is more than you need.

mostly because I didn't know about them. I know dlt and lto are good
technologies, at least if the drive is bought from a good
manufacturer. I do not know as much about every type of tape out
there.

the lto claims to be able to do the whole 200g in 2 hours. that is a
consideration. this is not a 24hour operation, but considering they do
work into the late evening, and certain automated EDI/import/export
procedures start at 4:45 am, and that the users are scattered from
coast to coast, that all adds up to there being not very many
regulalrly dead hours at night. (between 8-9pm on west coast, and
4:45am east coast)


> 
> How big and how many disks are in the machine natively.
> 
> >Using disks, they can just buy bigger disks. Even now, disks are
> >available almost twice as large as all their raid arrays put
> >together.
> 
> And how large is their raid array?

at the moment, a raid-10 (yes 10) array of 4 36 gig drives, yeilding
73 gigs of useable space, plus a raid-0 array of two 73 gig drives
yeilding another 73 gigs. the raid10 is for the / & /u (the OS &
database application) the raid0 is for the scanned documents. probably
the raid10 will not need to be increased for a few years, the raid0
may need to be increased in as little as a year, but probably more
like almost 2.

this is a new server so at the moment they are only using about 50% of
everything (about evenly split too) hmm... they are already at 60% of
the raid0 for scanned documents. may need to recalculate my timetable
after all :)

the plan was to get bigger drives after a year or so when they are
cheaper, and the 140 is big enough to get them that far easily.
getting 6 drives all double the current sizes is kind of expensive
considering they just bought a new dual-p4-xeon 2.2ghz, 66MHz 64bit
pci raid card, etc, the $6000 network backup unit, paid us a bunch of
on-site hours for me to do the install & migration, plus travel &
hotel etc... can't do *everything* you would like to all at once
sometimes, so the drive space is not spectacular at first, it's enough
to go until drives are cheaper and I don't even have to go on site to
migrate to the new drives when the day comes.
 

Comments?

Jump to Comments

Many of the products and books I review are things I purchased for my own use. Some were given to me specifically for the purpose of reviewing them. I resell or can earn commissions from the sale of some of these items. Links within these pages may be affiliate links that pay me for referring you to them. That's mostly insignificant amounts of money; whenever it is not I have made my relationship plain. I also may own stock in companies mentioned here. If you have any question, please do feel free to contact me.

Specific links that take you to pages that allow you to purchase the item I reviewed are very likely to pay me a commission. Many of the books I review were given to me by the publishers specifically for the purpose of writing a review. These gifts and referral fees do not affect my opinions; I often give bad reviews anyway.

We use Google third-party advertising companies to serve ads when you visit our website. These companies may use information (not including your name, address, email address, or telephone number) about your visits to this and other websites in order to provide advertisements about goods and services of interest to you. If you would like more information about this practice and to know your choices about not having this information used by these companies, click here.