On Mon, 10 Nov 2014 09:02:07 -0800, John Larkin Gave us:
Even better would be to use a full RAID system with 5 or more drives, and you could do your backups onto mSATA SSDs, which take up about 1
40th the volume of a 3.5 inch platter based drive, AND they are way faster too.
At the end of it, you have a bunch of really good SSDs to fashion your next RAID array out of. Eventually, you'll have a RAID array wherever you keep or generate sensitive data, and you would eventually make you back ups onto RAID arrays as well. Get a nice, deep floor safe (installed into a slab) and you do not have to worry about "off site" requisites any more either. The drive form factors allow easy use of the safe. No more lugging drives from place to place.
Smaller SSDs are cheaper, so a RAID array of them is a cheaper build, and results in a more reliable data store, AND it works nearly an order of magnitude faster, so even makes all those SPICE guys happy, even though they do not know where the bottleneck is in that process either..
I have a couple of 3-TB RAID-5 NAS boxes, one at the lab and one at home.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510
hobbs at electrooptical dot net
http://electrooptical.net
But DVD's and CD's don't die arbitrarily like USB drives... I had one die on me last month. Fortunately it was just stuff of which I had original copy. ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
All my three "work" PCs have front-panel hot-plug RAID drives. Never had a failure. What's cool is that I can clone my work PC OS to my home and cabin PCs easily. And I can occasionally drop off a full drive as a spare/backup/checkpoint, with everything installed and ready to run if needed.
I think my Spice runs are compute limited. Possibly slowed down by logging sim data to disk. I'm seeing about 55% CPU load by LT Spice on my switcher sim. So maybe a non-RAID SSD would help there, for temp storage of the sim data.
--
John Larkin Highland Technology, Inc
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
My backup problem isn't losing copies, it's having too many!
We do multiple rolling hard-drive backups of all our official company files. And once a month, it gets written to a metal-tube-O-ring USB memory stick, with copies of that all over California.
--
John Larkin Highland Technology, Inc
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
Mine are Synology DS411slim, which was recently discontinued but is still in support. I basically just use it for local NetBIOS and remote SSH/SFTP. The discs are HGST 1.5 TB 2.5-inch, and have been running for six months or so, non-stop. I had one infant-mortality failure, but the other seven have been working flawlessly (zero restarts on S.M.A.R.T).
I have both of them on IBM-rebadged APS SmartUPS 750 UPSes that I got on eBay for about $50 each.
The NAS knows how to talk to the UPS to make sure it shuts down cleanly in the event of a long power outage.
But I still take DVD images occasionally.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510
hobbs at electrooptical dot net
http://electrooptical.net
I live in an air-conditioned world with exceptionally low humidity, and each CD/DVD is saved in its own container so they can't rub and scratch... so I have CD's dating back decades that read just fine.
Some years ago, when I still had a 5-1/4" floppy drive that a PC knew how to talk to, I copied all my really old stuff to CD's ;-) ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
On Mon, 10 Nov 2014 11:38:53 -0700, Jim Thompson Gave us:
Good move, because floppies are notorious for becoming unreadable. And it is usually the drive, not the disc. Picky read/reject/give-up code. My 2.88 laser aligned floppies NEVER EVER lost a single bit of data ever. Try to find a drive or motherboard BIOS with support though. Near impossible. Sad too, since they were far superior. Got left behind with other "overpriced" (by the mindset of the day) IBM PS2 type "proprietary think" technology, just as they ushered in Iomega's triple overpriced utter crap, which everyone embraced simply because of capacity numbers. I still have 2.88 drives and discs and even an old machine somewhere with the right BIOS. I think one can still buy a PCIs I/O card with a floppy interface on it that carries it.
I wanted to backup my DesqViewX install floppy images, and was so frustrated trying. Then I found the dang image set online! AND QEMM
386 too! Pretty sure even Bloggs could hunt up and DL a full set within a few minutes online, if he wasn't such a caveman.
I now look online for old legacy stuff, as trying to round up working install packages here is not always easy. No nefarious purpose. I own the discs, so finding already done copies is not a crime. Gotta love the internet. AND QEMU AND DOSBox.
AND Old hardware! (I speak for myself :-)) (where's my coffee?!)(Maybe put some Jamison's in there)
I still (very) occasionally find a use for QEMM386.
About the only legacy stuff I use now are a tool that converts PCWrite document files to Word; and I, once in a while, haul out the old original OrCAD symbol libraries so I can convert those schematics to PSpice.
Sounds good... I'm a Drambuie fan myself ;-) ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...
As long as they don't age like 8 inch floppy disk. In 25 years, will you be able to read them on your current hardware or will you need to keep legacy devices around?
I kept an old WFWG machine around for many years, hung on the network, just so I could read old floppies. When it finally started doing weird crashes I quickly read off everything from 5-1/4" and 3-1/2" onto CD's.
When I see signs of new equipment not supporting CD's I'll do a similar transformation. (I have schematics saved from at least 35 years back on CD's... before that, paper, though I have converted even some of those to PDF's ;-) ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Until, of course, you get a "single spindle failure" and discover you now have terabytes of data that need to be recovered by a professional service bureau!
(I keep all my "precious" archives on small spindles in redundant configurations. My archives date back more than 30 years and I've not lost anything -- yet! Wanna know how many minutes I spent talking on the phone with client X in 1985 and what each call was about?? :-/ )
As with any mechanism upon which you presumably rely, you *have*, of course, tried "dropping" a drive and rebuilding the array? I.e., you (presumably) TRUST the array so shouldn't feel nervous about pulling a single drive and watching to see how it fares? (after all, even if one of the remaining drives COINCIDENTALLY died, you'd still have the HOPEFULLY good drive in your hands with which to rebuild the array!)
With consumer-grade drives, once you get to larger capacities, a *real* failure often results in an inability to rebuild the array before some
*other* aspect of the array goes down (e.g., another drive that is exactly the same age/make/model as the first failing drive, has lived in the same environment as the first failed drive, is powered by the same power supply, driven by the same processor, etc.). You might be surprised at just how long it takes to rebuild a large array! Keeping in mind that you have *no* protection during that process (and, no guarantee that the array *will* be rebuildable -- an unrecoverable read error on one of the remaining drives can leave you with !)
If your NAS doesn't contain a hot spare, then you have to hope you "notice" the failure and take actions to recover before the other shoe drops...
Consumer-grade NAS's are another can of worms altogether!
Have you tried pulling the *set* of drives and installing them in another ("identical") RAID enclosure to see if they will be recognized as a complete set or (despairingly!) marked as "foreign" (and, thus, trash!)? Said another way, when your "box" dies, how will you access/recover the data that it was charged with protecting?
When my archive started growing, I initially adopted the approach of running shelves -- first as JBOD's, then escalating levels of RAID. But, it was damn near impossible to *think* in the same room (a dozen or more 15K U320 drives)! Way overkill (and, too many single points of failure!) Yet, this was all enterprise class hardware!
If you want to keep a single archive in a single enclosure, then do yourself a favor and step up to RAID6 -- or over to RAID10 -- or reduce the size of the archive.
Or, plan on a preventative replication and replacement strategy.
If it's truly an *archive*, then the biggest favor you can do yourself is to move to redundant *boxes* so they see different environments, wear, etc.
I currently use multiple small spindles on different hosts and different media types. I've abandoned the large shelf approaches (though the redundant power supplies gave some level of reassurance) and am now moving towards
*individual* drives in USB enclosures. This gives you the ability to swap drives (even "hot"!) without having to keep a boatload of sleds for different shelves. It also lets you mix and match drive technologies (to a limited extent -- but, far more than a shelf or dedicated appliance does!)
And, if you adopt the goal of making media "portable" between "machines" (whether those are NAS's, workstations or whatever), then even a failure of "The NAS" on which a volume was mounted is only a minor inconvenience... just unplug the USB cable and plug it into another NAS, Workstation, etc.
I'd have loved to be able to find a "single drive NAS" onto which I could install my own software and, thus, let the "drive interface" be a network protocol instead of "USB protocol". But, the price point of processor+power supply+medium just doesn't make sense in today's market!
So, I'll be assembling multi-drive (USB) NAS appliances out of tiny (SBC) headless machines. That way, I can mount 1 drive (or 20!) concurrently as my needs dictate. Or, just SneakerNet individual drives to whichever workstation needs "lots of access" to it! (I used to do this with off-lined SCSI/SCA drives -- plug drive into enclosure and spin it up for "local" access!)
[Of course, this LOW level of performance is satisfactory as it's JUST an archive -- not a file service -- and only has to support a single user! But, even USB2 drives can EASILY saturate 100Mb links (though my workstations and servers are all Gb)]
Remove one of your drives (or replace it with a foreign drive) to simulate a failure. See how long it takes to rebuild the array.
With large drives (failing "naturally"), the time it takes for an array rebuild -- plus the likelihood that you will encounter an unrecoverable read error on one of the other drives in the set -- leads to alarmingly high *practical* failure rates.
Here's one such calculator to give you an idea how *often* you can expect to encounter a failure. And, how often you can expect to LOSE DATA! :>
On my desktop PCs (ProLiants with hot-plug RAID) I occasionally pull out a drive for one reason or another, and poke in a replacement. It takes about 1.5 hours to resync them.
We've used floppies, tape, hard drive cartriges, CDs, DVDs, and USB sticks for backup. I don't think we've ever lost anything important.
Just make lots of copies.
--
John Larkin Highland Technology, Inc
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
On Mon, 10 Nov 2014 22:10:50 -0600, Les Cargill Gave us:
snip
That's just silly. The file copy code, and the drive hardware itself should insure that each file was written correctly. This extra step just further exercises the equipment, narrowing the window until a possible failure. With the backup being at least mirrored, there is yet another check going on. No need to worry, Les. They spent decades incorporating that into the drive itself.
Such checks are for optical disc burn efficacy proofing, not much else. Really big datagrams. Huge database files.
Hard to write a bad file IF files are what is being backed up. Another reason that "backup without an OS" thread was doomed.
It all depends on who you trust here - your server must be with someone you trust (you are also trusting them with the privacy of your mail, the reliability of your mail transfers, and with the continuity of your mail address). If you are not keen on Gmail, that's fine - certainly a zero-cost provider has fewer responsibilities than one who charges for their services. At my company, the mail server is at a company we trust to handle backups, redundancy, etc. My personal mail is on a server that I trust - it is in my cellar at home (a dovecot imap server is easy to get in place, and requires no maintenance).
But all this is a side issue - as I said above, client-side backup of IMAP is no harder than client-side backup of POP3. With IMAP, you can easily have the same backups to discs, /and/ you (should) have backups at the server.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.