dufffader
Leicanaut/Nikonaut...
Last weekend I had to move to a new rental apartment in the same building. In the process of moving, the primary HDD on my home server crashed (it was an Acer H340) and I had problems booting it up once I took it out of the box. I kept my cool, thinking it couldn't be that bad as I have the data duplicated, not exactly RAID, but close, Windows Server-style.
What do you know, spent the next few days trying to recover the primary HDD and install the WHS system software without success. Which means no way to read all those HDD that are attached to it. That PITA Acer box just kept on giving me all sort of LED light colour combination and network error when I tried to install a new OS onto the primary partition. Online server forums was not helping, with stories of guys losing data while recovering and instructions only a programmer would understand.
I do my photography in both film and digital. Generally, grab shot and landscape photos are captured on a digital camera while street photos are all on film RF and SLR cameras and then scanned. It was during these few days when I thought about what would happened if I lost my two decade of photos, initially thinking that a server with duplication was safe enough. I thought about the time it would have taken to rescan the hundreds of rolls of film in the home server (that would have taken probably half a year or more, in my spare time). Obviously the DSLR and digital point & shoot photos would be gone forever. And for a while I did think I would have given up photography if I couldn't recover all my past work from this stinking piece of hardware that's just not cooperating. I don't think I would have the energy to get out of that rut and restart everything all over again. Maybe pick up some other hobby, like racing cars or womanising!
Anyway, I'm here now waiting for a new NAS to arrive in the post. And I think I might have found a way to recover the information as the other 3 drives has some kind of data left that was readable when I plugged them into an external USB drive. On top of that, I was glad I have a large HDD of the server backup that I do monthly sitting in my office (off site storage), although it would still require me to scan the last few week's work off negatives.
I think I should get out of this situation fine, but what a thought...
What do you know, spent the next few days trying to recover the primary HDD and install the WHS system software without success. Which means no way to read all those HDD that are attached to it. That PITA Acer box just kept on giving me all sort of LED light colour combination and network error when I tried to install a new OS onto the primary partition. Online server forums was not helping, with stories of guys losing data while recovering and instructions only a programmer would understand.
I do my photography in both film and digital. Generally, grab shot and landscape photos are captured on a digital camera while street photos are all on film RF and SLR cameras and then scanned. It was during these few days when I thought about what would happened if I lost my two decade of photos, initially thinking that a server with duplication was safe enough. I thought about the time it would have taken to rescan the hundreds of rolls of film in the home server (that would have taken probably half a year or more, in my spare time). Obviously the DSLR and digital point & shoot photos would be gone forever. And for a while I did think I would have given up photography if I couldn't recover all my past work from this stinking piece of hardware that's just not cooperating. I don't think I would have the energy to get out of that rut and restart everything all over again. Maybe pick up some other hobby, like racing cars or womanising!
Anyway, I'm here now waiting for a new NAS to arrive in the post. And I think I might have found a way to recover the information as the other 3 drives has some kind of data left that was readable when I plugged them into an external USB drive. On top of that, I was glad I have a large HDD of the server backup that I do monthly sitting in my office (off site storage), although it would still require me to scan the last few week's work off negatives.
I think I should get out of this situation fine, but what a thought...
If the data is very important, and you are not technically competent enough to maintain and retrieve your backup, it would be a very good idea to pay someone else to do it.
Maybe a programmer who "understands it" is the right person for this job.
Maybe a programmer who "understands it" is the right person for this job.
greyelm
Malcolm
Have you heard of Spinrite? I've heard many good things about it.
http://www.grc.com/sr/spinrite.htm
http://www.grc.com/sr/spinrite.htm
Shutterspark
The perpetual new guy.
Hope it works out for you in the end.
I keep all my photos backed up online as well as locally. Best solution IMO.
If you don't need raw then just get a Pro account at Flickr or some other large and well known photo sharing site and let them handle the expensive redundancy backups.
And there are plenty of more serious cloud backup solutions which insure against data loss. It usually works out cheaper than having that kind of reliability in your own house. More convenient too, assuming you have a fairly good connection.
I keep all my photos backed up online as well as locally. Best solution IMO.
If you don't need raw then just get a Pro account at Flickr or some other large and well known photo sharing site and let them handle the expensive redundancy backups.
And there are plenty of more serious cloud backup solutions which insure against data loss. It usually works out cheaper than having that kind of reliability in your own house. More convenient too, assuming you have a fairly good connection.
dufffader
Leicanaut/Nikonaut...
They are separate drives but part of a windows home server package, which means it appears as one big drive to the computer and the server does all the background plumbing, such as data duplication across the drives. The issue is, without the primary drive, the 3 other drives has the data spread across them. And you were right, they are readable, but has to be assembled manually since the database that assembles them together is on the primary drive. This is windows home server v1 with drive extender. Well, I only figured out they are readable this morning. On a Mac the files are all hidden and incomplete.
I wouldn't say I'm computer illiterate, but with some redundant storage, I think there may be cases where the data may not be completely recoverable, unless one has a backup of the backup.
I wouldn't say I'm computer illiterate, but with some redundant storage, I think there may be cases where the data may not be completely recoverable, unless one has a backup of the backup.
dufffader
Leicanaut/Nikonaut...
I've heard of them. Even listen regularly to Steve's weekly podcast. I think I'll be able to get through this without spinrite. The data is there somewhere.
flip
良かったね！
Never lose hope on hard drive failure. It's usually just a question of time, money, creativity and/or expertise to retrieve it. Best of luck.
My wild guess as to what he's talking about is that he has some sort of a multi-drive RAID system, but the RAID configuration software and/or drivers are on the crashed drive, therefore he can't reconstruct the RAID to retrieve his data.
Richard G
Mentor
It also reflects how tight and redundant and up to date a backup strategy has to be for it not to cause days of time consuming fiddling, loss of a few days' data, possibly very important data, and general anxiety and frustration. People who run backup services for a living can have a catastrophe like a fire damaging a client's main office or one of the backup sites, and within an hour have an alternative site running the main business or backup server - and do it by telephone from anywhere in the world, by phone, within an hour or two.
Simulation is one of the steps in checking a backup strategy - preferably actually retrieving a backup and reconstituting a functioning system, but at least seriously analyzing the consequences of each potential breakdown and what would actually be required to retrieve the situation. Nearly always failures of hardware lead to lots of time being required. My business could be up and running within 30-60 minutes of any catastrophic failure of the main server and it's main backup. The cost of such a quick turn around is not much as part of the solution is having a very simple multiply reedundant backup of the crucial files, as well as already having a satisfactorily specced machine that could be the backup to the damaged server, just in case. I use Mac minis for this and it has been a trouble free strategy for six years, tested for real one evening two years ago and we were up and running again in 45 minutes.
Simulation is one of the steps in checking a backup strategy - preferably actually retrieving a backup and reconstituting a functioning system, but at least seriously analyzing the consequences of each potential breakdown and what would actually be required to retrieve the situation. Nearly always failures of hardware lead to lots of time being required. My business could be up and running within 30-60 minutes of any catastrophic failure of the main server and it's main backup. The cost of such a quick turn around is not much as part of the solution is having a very simple multiply reedundant backup of the crucial files, as well as already having a satisfactorily specced machine that could be the backup to the damaged server, just in case. I use Mac minis for this and it has been a trouble free strategy for six years, tested for real one evening two years ago and we were up and running again in 45 minutes.
dufffader
Leicanaut/Nikonaut...
You're right here. I bought into the "this product will solve your problem forever" storyline and should have simulated at least a couple of worst case scenario.
craygc
Well-known
I spend a lot of time and energy on backup strategies so while this may not address the current dilemma, its worth thinking about moving forward.
Actually, on your current dilemma, I did a little research on the Acer H340 box and found this:
No matter what you do, as mentioned above, nothing beats simulation for testing backup strategies. However, a good objective think through what happens if each element fails is a good starting point:
Single drives connected via USB or Firewire usually pose no issues, apart from a single drive failure. Connect or cycle two or more and you're more or less ok as the each drive is based on the host OS file system.
NAS boxes are a different story though. RAID or not to RAID (as in the case of the Acer box) is often brought up. Firstly no RAID config represents a backup strategy - its ONLY about hardware redundancy. The bigger issue to understand with a NAS is how is the information written to the drive - and this is a reasonably technical question but an important one.
In the case of the Acer H340 there is no RAID but rather a proprietary controller that managers the data. It doesn't offer any redundancy, more of a JBOD config (Just a Bunch Of Disks) managed and made to look like one but, by definition of having the controller partition on a primary drive, offers a higher rate of failure risk than RAID or a single disk. Just hope that recovery disk is available and works.
Most RAID based NAS these days are Linux boxes and for the main, that shouldn't have any interest to you. You connect via Ethernet and that's it. However, non-drive system failures can be catastrophic if the wrong configuration is used. RAID 5 and 6 is frequently used these day as it maximises disk usage and allows 1 or 2 disk failures out of a 4 or 5 disk cluster with full recovery being possible. That recovery is usually only possible is the NAS is fully operational.
This leads to considering a failure of the NAS hardware. RAID 5 and 6 is almost always based on a proprietary implementation, which in itself is not a concern but what do you do if all your drives in the RAID volume are good but you NAS is dead? I frequently see, even for top of the line SOHO NAS boxes, that newer models don't support previous model RAID volumes; ie. you cant even buy a new "same brand" NAS and reinstall all your drives - and you can guarantee that by the time you have a system failure, your NAS model is out of production. Without that proprietary controller you are going to have an immense amount of trouble recovering your data.
The reason RAID 5 and 6 are popular is that Enterprise deploy it heavily. However, they use Unix or Windows based servers where both hardware and software are standards along with full backup strategies. For any failure, drives can be fitted back into a new server immediately and/or rolled back to a checkpoint based on the last backup. Not something people do at home.
So for my recommendations for SOHO NAS boxes - use RAID 1 only - ie. 2 drives with data replicated on both, Also research to understand the file system used on the drives. As I mentioned above, most RAID capable NAS use a version of Linux and the file system is most likely something called EXT3. With easily available drivers for both Windows and OS-X to support EXT3 file systems, recovery is as easy as pulling one of the drives from the NAS, putting it into an external drive cover and connecting it via USB.
As a true anecdotal story, back in early June I had a power supply failure on my NAS; not so bad but it was out of production and the box obviously stopped working. A week later I lost my job and guess where my CV was? On the NAS. I had found a replacement power supply but it was going to take few weeks to ship in. However, all I needed to do was remove one of the RAID 1 drives and plug it into a Firewire based bare-drive housing I usually run time-machine on, and immediately I had full access to the disk. By the way, my backup was on the other RAID 1 in the same NAS
Given the above option for recovery, if you're using a typical 4-bay, or greater, NAS use another drive or a second RAID 1 volume to backup the primary RAID 1 to - this can be supplemented or replaced by external drives or DVDs as well. I have from that Goodsync software is reasonably cheap and offers a lot of control for ensuring multiple drives are then sync'ed accurately and regularly.
A bit long winded but hopefully helpful...
Actually, on your current dilemma, I did a little research on the Acer H340 box and found this:
no doubt you've just learn't this the hard way, however...
Looks like you may have a recovery option
No matter what you do, as mentioned above, nothing beats simulation for testing backup strategies. However, a good objective think through what happens if each element fails is a good starting point:
- Hard disk
- The box housing the hard disk
- Power supply
- etc...
Single drives connected via USB or Firewire usually pose no issues, apart from a single drive failure. Connect or cycle two or more and you're more or less ok as the each drive is based on the host OS file system.
NAS boxes are a different story though. RAID or not to RAID (as in the case of the Acer box) is often brought up. Firstly no RAID config represents a backup strategy - its ONLY about hardware redundancy. The bigger issue to understand with a NAS is how is the information written to the drive - and this is a reasonably technical question but an important one.
In the case of the Acer H340 there is no RAID but rather a proprietary controller that managers the data. It doesn't offer any redundancy, more of a JBOD config (Just a Bunch Of Disks) managed and made to look like one but, by definition of having the controller partition on a primary drive, offers a higher rate of failure risk than RAID or a single disk. Just hope that recovery disk is available and works.
Most RAID based NAS these days are Linux boxes and for the main, that shouldn't have any interest to you. You connect via Ethernet and that's it. However, non-drive system failures can be catastrophic if the wrong configuration is used. RAID 5 and 6 is frequently used these day as it maximises disk usage and allows 1 or 2 disk failures out of a 4 or 5 disk cluster with full recovery being possible. That recovery is usually only possible is the NAS is fully operational.
This leads to considering a failure of the NAS hardware. RAID 5 and 6 is almost always based on a proprietary implementation, which in itself is not a concern but what do you do if all your drives in the RAID volume are good but you NAS is dead? I frequently see, even for top of the line SOHO NAS boxes, that newer models don't support previous model RAID volumes; ie. you cant even buy a new "same brand" NAS and reinstall all your drives - and you can guarantee that by the time you have a system failure, your NAS model is out of production. Without that proprietary controller you are going to have an immense amount of trouble recovering your data.
The reason RAID 5 and 6 are popular is that Enterprise deploy it heavily. However, they use Unix or Windows based servers where both hardware and software are standards along with full backup strategies. For any failure, drives can be fitted back into a new server immediately and/or rolled back to a checkpoint based on the last backup. Not something people do at home.
So for my recommendations for SOHO NAS boxes - use RAID 1 only - ie. 2 drives with data replicated on both, Also research to understand the file system used on the drives. As I mentioned above, most RAID capable NAS use a version of Linux and the file system is most likely something called EXT3. With easily available drivers for both Windows and OS-X to support EXT3 file systems, recovery is as easy as pulling one of the drives from the NAS, putting it into an external drive cover and connecting it via USB.
As a true anecdotal story, back in early June I had a power supply failure on my NAS; not so bad but it was out of production and the box obviously stopped working. A week later I lost my job and guess where my CV was? On the NAS. I had found a replacement power supply but it was going to take few weeks to ship in. However, all I needed to do was remove one of the RAID 1 drives and plug it into a Firewire based bare-drive housing I usually run time-machine on, and immediately I had full access to the disk. By the way, my backup was on the other RAID 1 in the same NAS
Given the above option for recovery, if you're using a typical 4-bay, or greater, NAS use another drive or a second RAID 1 volume to backup the primary RAID 1 to - this can be supplemented or replaced by external drives or DVDs as well. I have from that Goodsync software is reasonably cheap and offers a lot of control for ensuring multiple drives are then sync'ed accurately and regularly.
A bit long winded but hopefully helpful...
ath
Well-known
My backup strategy is simply and IMHO reliable:
I just copy "My Documents" to three external Harddisks. These are from different manufactureres and not connected to the PC and/or mains when not in use for backup.
One is FAT32, the others NTFS.
Just plug the HD in the next PC and everything is there.
Schedule for backup is at least once a week plus after significant changes.
I make sure that "My Documents" contains all important data; for some programs (e.g. email, Book Design SW) you have to assign a home folder within "My Documents". And I keep a copy of all program installation files there as well.
The boot disk with installed programs is cloned to another HD.
When the data disk of my PC developed an error it was a matter of 5 minutes and I was up working again.
I just copy "My Documents" to three external Harddisks. These are from different manufactureres and not connected to the PC and/or mains when not in use for backup.
One is FAT32, the others NTFS.
Just plug the HD in the next PC and everything is there.
Schedule for backup is at least once a week plus after significant changes.
I make sure that "My Documents" contains all important data; for some programs (e.g. email, Book Design SW) you have to assign a home folder within "My Documents". And I keep a copy of all program installation files there as well.
The boot disk with installed programs is cloned to another HD.
When the data disk of my PC developed an error it was a matter of 5 minutes and I was up working again.