This idea arises since IDE hard disk nowadays are becoming bigger and cheaper. My plan is to use an existing P4 512MB PC and install SATA IDE 160 GB mirroring (with SATA RAID 1 Card) and use instead of buying External SDLT Tape 110/220 drive with 10 Tape cartridges and Veritas Backup exec 9. I can add couple more 160 GB on IDE ports for added redundancy or disregard the RAID and slot-in 3 160 GB hdd. Can anybody advice the pros and the cons, cost and performance? Thanks in advance for the input.
Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations SkipVought on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
Alternative for External SDLT Tape drive 110/220 GB 2
- Thread starter ricpinto
- Start date
NickFerrar
MIS
It depends what you do with your backups, we store ours off-site so they're available if there's a disaster situation at the office. If you're doing disk backups rather than tape backups then you lose much of the point of backups IMO.
-
1
- Thread starter
- #3
Let us know how this turns out, the idea had crossed my mind as well. It seems cheaper to pick up the new PC, drives and card than a full tape drive route. Licensing may play into it, but for ease of use it would be nice.
Is the DR functionality on the software any good?
- Thread starter
- #5
My DR testing is on hold as the trial version of
does not allow DR functionality
. I'm hesitant to buy without testing it first. Is there anybody out who have already tried this software?
does not allow DR functionality
. I'm hesitant to buy without testing it first. Is there anybody out who have already tried this software?I'd recommend using a combination of tape and RAID solutions, but you really need to determine what's more important. If uptime is more important, use a RAID setup. If keeping your data safe from harm is more important, use tapes. I want the best of both worlds so I do either RAID1 or RAID5 in addition to nightly tape backups.
RAID is good at recovering from drive failures but can't help much if just the data has been tampered with. The employees can continue to work but usually at a slower pace while the data is resynched or rebuilt across the array. However, drive prices are coming down while their speed and storage abilities are getting better. On the bad side, it's somewhat difficult to protect the data from physical abuse like fire since it's usually a bit more difficult to move the drive to safety in time.
Tapes are good for recovering from bad or deleted data with relatively minor losses if backups are done on a daily basis. Also, they're easily stored off-site or in something like a fireproof filing cabinet in the event that disaster strikes. However, recovery time can take a while and it may put your business at a stand-still while the server is offline during the restore procedure. Unfortunately, a tape system is definitely going to cost you more in the long run. I just spent about $2500.00 for a 6-tape autoloading setup that can store about 432GB, max. I could have bought a few high capacity hard drives for that much money.
Like I said before, you really just have to ask yourself what's more important and go from there.
RAID is good at recovering from drive failures but can't help much if just the data has been tampered with. The employees can continue to work but usually at a slower pace while the data is resynched or rebuilt across the array. However, drive prices are coming down while their speed and storage abilities are getting better. On the bad side, it's somewhat difficult to protect the data from physical abuse like fire since it's usually a bit more difficult to move the drive to safety in time.
Tapes are good for recovering from bad or deleted data with relatively minor losses if backups are done on a daily basis. Also, they're easily stored off-site or in something like a fireproof filing cabinet in the event that disaster strikes. However, recovery time can take a while and it may put your business at a stand-still while the server is offline during the restore procedure. Unfortunately, a tape system is definitely going to cost you more in the long run. I just spent about $2500.00 for a 6-tape autoloading setup that can store about 432GB, max. I could have bought a few high capacity hard drives for that much money.
Like I said before, you really just have to ask yourself what's more important and go from there.
- Thread starter
- #7
is more of imaging the server harddisk realtime(with scheduler) and whole-volume or file-by-file recovery, also realtime. I've used RAID 1 on the PC for added data protection. Actually I have 40/80 DLT on each server but since we are planning to buy a new server with hundreds of gig in it then it must have a Super DLT. So the idea arises, I don't recommend this setup alone, it must run in complimentary with a tape drive(image from powequest must be written on a tape for offsite safekeeping), most probably I will use one of the 40/80 DLTs for that purpose or a CD writer(the program supports creating several 650 mbs to make 1 image for burnning it in a CD-RW). If this one materialise, I will image all the system/boot partition of all the servers for faster DR.
We use external USB 2.0 IDE hard drives instead of tapes for our data backups which can be removed off-site for extra protection. Much faster and less expensive than any tape drive I've ever tried. Also, depending on your choice of backup software, the recovery is usually a lot more convenient than tapes, especially for recovering only individual data files.
*Shakes head as he realizes he posted when half asleep*
I guess I just blew right past the PowerQuest link you had. I went and looked at it, and it seems pretty cool. If I could, I'd return the new TBU and buy the PowerQuest software and some high capacity drives. For me, the cost would be about the same (including the software and some high capacity drive equipment), but it would definitely save me a lot of time and trouble in the event of a problem. It would probably last a whole lot longer as well. Over a 3-5 year period, you'd probably spend a lot less money overall just because you wouldn't have to keep replacing worn-out tapes.
I'll have to keep this in mind for the next time I need to go in search of a backup solution.
I guess I just blew right past the PowerQuest link you had. I went and looked at it, and it seems pretty cool. If I could, I'd return the new TBU and buy the PowerQuest software and some high capacity drives. For me, the cost would be about the same (including the software and some high capacity drive equipment), but it would definitely save me a lot of time and trouble in the event of a problem. It would probably last a whole lot longer as well. Over a 3-5 year period, you'd probably spend a lot less money overall just because you wouldn't have to keep replacing worn-out tapes.
I'll have to keep this in mind for the next time I need to go in search of a backup solution.
-
1
- #10
A couple points from my experience:
1. You'll probably want RAID on all of your servers to meet uptime requirements. I've had several hard disk failures in the past year and never had to take a server down thanks to RAID 5. It works great for fault tolerance, but it does nothing for disaster recovery and as such doesn't really work into a proper backup strategy.
2. Most companies backup to tape to allow offsite storage of data. While it's true that using external IDE drives you can store the drive offsite, you'll run into a couple of other problems.
The first problem is storage ability. If your off-site storage company is only set up to store tapes, it will be considerably more expensive to store external disk units with them offsite.
The second problem is reliability. Hard disks are fairly fragile, especially compared to tapes. It would be very easy for an external disk to be damaged in transit, rendering the data on it useless.
The third problem is cost. While it's true that a couple of external IDE units are cheaper than a tape drive and blank tapes, it gets more expensive over time. Don't think about the 1-time cost of a tape drive when trying to figure out what your media costs for the year are, it will skew the math.
Most companies with a 4-week rotation using incremental and full backups with backups transported offsite daily will have a minimum 28 tapes for that cycle. Replacing those tapes with 28 external drive units will be costly, and that doesn't even address media for monthly or annual archival backups. A single 100/200 GB Ultrium tape costs about $45, less if you buy in bulk. That's hard to beat when you're comparing to external drive units.
For example, at my company we are using 76 tapes per year to back up a specific set of servers, and that rotation allows us to restore our system to the state that it was in at any specific date during the year. That works out to around $3000 per year in media costs (we get a discount from buying bulk). If we replaced those tapes with external drives, it would cost many times that amount.
3. Hard disk space is cheap. You can certainly find ways to include hard disks into your backup solution to improve recovery performance. If you had a drive array/SAN/NAS device on your network that is dedicated to backups, you could store a second copy of each backup on the array. In the event that you needed files restored from the backup, you could immediately pull them from the disk array in the server room instead of having to order a tape delivery from your offsite provider. This would generally be much faster, and your users would love you for it. More and more companies are doing this today since hard disk space is so cheap now.
We actually looked at this initially when we worked out our current backup strategy. The problem that we had is that we're generating around 220 GB of backup data every week. Since most requests for file restores are for files that were backed up in the previous month, we needed a storage solution that would provide approximately 1 TB of data. While that was certainly doable, we found the cost to be prohibitive. Under normal circumstances, we actually only end up doing file restores about once a month, usually because someone deleted a file by accident. I couldn't justify spending another $13000 just to be able to recover a deleted spreadsheet in 10 minutes versus 4 hours. Especially since most restores aren't so urgent that you couldn't wait a few hours.
My advice to you is to read a few whitepapers on backup strategies for disaster recovery (Veritas has some online). Figure out what specifically you need to be able to do. Figure out what kind of cycle will be required to support the capabilities that you need. Then determine which media format is most economical for you while allowing for some growth).
1. You'll probably want RAID on all of your servers to meet uptime requirements. I've had several hard disk failures in the past year and never had to take a server down thanks to RAID 5. It works great for fault tolerance, but it does nothing for disaster recovery and as such doesn't really work into a proper backup strategy.
2. Most companies backup to tape to allow offsite storage of data. While it's true that using external IDE drives you can store the drive offsite, you'll run into a couple of other problems.
The first problem is storage ability. If your off-site storage company is only set up to store tapes, it will be considerably more expensive to store external disk units with them offsite.
The second problem is reliability. Hard disks are fairly fragile, especially compared to tapes. It would be very easy for an external disk to be damaged in transit, rendering the data on it useless.
The third problem is cost. While it's true that a couple of external IDE units are cheaper than a tape drive and blank tapes, it gets more expensive over time. Don't think about the 1-time cost of a tape drive when trying to figure out what your media costs for the year are, it will skew the math.
Most companies with a 4-week rotation using incremental and full backups with backups transported offsite daily will have a minimum 28 tapes for that cycle. Replacing those tapes with 28 external drive units will be costly, and that doesn't even address media for monthly or annual archival backups. A single 100/200 GB Ultrium tape costs about $45, less if you buy in bulk. That's hard to beat when you're comparing to external drive units.
For example, at my company we are using 76 tapes per year to back up a specific set of servers, and that rotation allows us to restore our system to the state that it was in at any specific date during the year. That works out to around $3000 per year in media costs (we get a discount from buying bulk). If we replaced those tapes with external drives, it would cost many times that amount.
3. Hard disk space is cheap. You can certainly find ways to include hard disks into your backup solution to improve recovery performance. If you had a drive array/SAN/NAS device on your network that is dedicated to backups, you could store a second copy of each backup on the array. In the event that you needed files restored from the backup, you could immediately pull them from the disk array in the server room instead of having to order a tape delivery from your offsite provider. This would generally be much faster, and your users would love you for it. More and more companies are doing this today since hard disk space is so cheap now.
We actually looked at this initially when we worked out our current backup strategy. The problem that we had is that we're generating around 220 GB of backup data every week. Since most requests for file restores are for files that were backed up in the previous month, we needed a storage solution that would provide approximately 1 TB of data. While that was certainly doable, we found the cost to be prohibitive. Under normal circumstances, we actually only end up doing file restores about once a month, usually because someone deleted a file by accident. I couldn't justify spending another $13000 just to be able to recover a deleted spreadsheet in 10 minutes versus 4 hours. Especially since most restores aren't so urgent that you couldn't wait a few hours.
My advice to you is to read a few whitepapers on backup strategies for disaster recovery (Veritas has some online). Figure out what specifically you need to be able to do. Figure out what kind of cycle will be required to support the capabilities that you need. Then determine which media format is most economical for you while allowing for some growth).
From a smaller business perspective...
Since we only have about 10 gb of data to back up daily (20 users with some commercial lab data), we should be able to store multiple restore points on each HD. This would allow us to get away with 2-4 external drives (100-200gb with casing), back up one night, take it offsite the next, bring night before lasts back, etc. If the software can handle doing incremental to a certain backup set (versus the whole drive) then we should be able to get even more leg out of the drives. Even our full backup (root drive, non-critical data) is only about 30gb so this might work.
A 40/80 gb tape drive is about 800-900 bought online (pricewatch, versus $2000 straight from quantum) and tapes cost about $30 since we can't really buy in bulk. We'd probably use about 10-15 tapes per major cycle ($300-$450) So the whole deal would cost us between $1100-1400.
If we go with powerquest, its about $800 for the small business version, plus $150 (pricewatch) for each 200 gb external firewire drive. So for the same relative price we could get 2-4 external drives that would easily hold multiple restore points.
This all depends on whether or not the software can backup to 1 drive multiple times (anyone know?). With the casings, the drives are relatively stable and without a car accident or the person having a major snafu, the drives should last quite a while (3-5 years?)
Thoughts? For a larger data size I definately agree that a blended approach including tapes is necessary, but for a smaller company on a small budget this seems like a feasible solution (if everything works of course)
JT
Since we only have about 10 gb of data to back up daily (20 users with some commercial lab data), we should be able to store multiple restore points on each HD. This would allow us to get away with 2-4 external drives (100-200gb with casing), back up one night, take it offsite the next, bring night before lasts back, etc. If the software can handle doing incremental to a certain backup set (versus the whole drive) then we should be able to get even more leg out of the drives. Even our full backup (root drive, non-critical data) is only about 30gb so this might work.
A 40/80 gb tape drive is about 800-900 bought online (pricewatch, versus $2000 straight from quantum) and tapes cost about $30 since we can't really buy in bulk. We'd probably use about 10-15 tapes per major cycle ($300-$450) So the whole deal would cost us between $1100-1400.
If we go with powerquest, its about $800 for the small business version, plus $150 (pricewatch) for each 200 gb external firewire drive. So for the same relative price we could get 2-4 external drives that would easily hold multiple restore points.
This all depends on whether or not the software can backup to 1 drive multiple times (anyone know?). With the casings, the drives are relatively stable and without a car accident or the person having a major snafu, the drives should last quite a while (3-5 years?)
Thoughts? For a larger data size I definately agree that a blended approach including tapes is necessary, but for a smaller company on a small budget this seems like a feasible solution (if everything works of course)
JT
I would still feel your concentrating too many "eggs" in toow few "baskets" with the disk strategies. They're great for rapid recovery, but I feel that's their only advantage right now.
I'm a small business also - 1 server, 15 workstations. I use tape - AIT-1. One advantage not mentioned here is archivability. I have a two week rotation of full backups, plus dedicated weekly and monthly tapes. With that, I've got 100% restore capability for any file/email within that two week window - as long as the file existed on disk at backup time.
I just don't see any disk based solutions out there in the smae price range as tape for small business that can give me that granularity of restore.
I'm a small business also - 1 server, 15 workstations. I use tape - AIT-1. One advantage not mentioned here is archivability. I have a two week rotation of full backups, plus dedicated weekly and monthly tapes. With that, I've got 100% restore capability for any file/email within that two week window - as long as the file existed on disk at backup time.
I just don't see any disk based solutions out there in the smae price range as tape for small business that can give me that granularity of restore.
I see your point, and raise you another:
True, but we are also talking about multiple drives each of which will contain a subset of the full backup scheme (every third or 4th day as well as every third Friday/month/6month etc). So if we have a drive failure we do have the other 2-3 offsite as backup to the backup.
There is the same benefit that if the server goes and they need immediate access it allows them to restore to a desktop if necessary. The offset is reinstalling the software versus re-installing the hardware tape drive.
we're actually in a situation right now where the tape drive has some really funky errors. We have a 20/40 HP colorado and every single tape gives the error "the media in the drive must be erased before it can be used" every time we try to restore anything. someone had mentioned that it may be a cleanliness issue, but the HP drive is supposed to be self cleaning and there aren't any cleaning tapes available (Travan). Every other piece of literature I read says that error only comes with a broken backup set, but this is also happening with single tapes.
So, with the tape drive "down" we're left with a lot of tapes with no way of getting to the data. Plus they're going to outgrow the 20 gb soon as they add more employees. So, in essence our basket full of eggs is currently out of reach. We do have a basic Raid setup for fault tolerance, but past date restores are unavailable right now.
And as a preference note, tapes are waaay slow and I would welcome the opportunity to attempt something different.
The immediacy of the disks is an advantage, but if I needed immediate access to deleted information or restore business, that soley tape or disk would be the way to go.
I've tested my recovery from a non-functional server status to a fully operational network and can get there in 2.5 hrs with my solution. That's comfortable for me. Given a complete destruction of my network, all I've got to do is come up with hardware, and that's doable within 24 hours given my established relationships and credit. If the business can't survive that window - then tape or disk isn't the solution.
Just as an aside - Travan is a tape technology that should have never been pushed to the enterprise level. The nice thing about the digital tape formats - DAT/AIT/DLT/LTO - is that they can be universally read - unless they've been encrypted.
I've tested my recovery from a non-functional server status to a fully operational network and can get there in 2.5 hrs with my solution. That's comfortable for me. Given a complete destruction of my network, all I've got to do is come up with hardware, and that's doable within 24 hours given my established relationships and credit. If the business can't survive that window - then tape or disk isn't the solution.
Just as an aside - Travan is a tape technology that should have never been pushed to the enterprise level. The nice thing about the digital tape formats - DAT/AIT/DLT/LTO - is that they can be universally read - unless they've been encrypted.
I'd like to quantify that by saying that Travan drives are slow, but not everything else is. Most tape drives are available with a speedy SCSI interface. In my example above, our full backup runs around 100GB and takes about 7 hours for a backup and full verify. This is across multiple servers and from a single media server, using a single Ultrium 100/200 drive. When you do your shopping, check to see what sort of throughput is possible with the various drives, as some are much faster than others.
The problem that you're having with your old solution may simply be that the tapes are worn out. Depending on the media used there are limits to the number of times that it can be overwritten (a lot of people miss that point).
You're right about having a rotation/subset of the full backup set on a single disk. That's the value of having multiple pieces of media no matter which format you choose.
I'd make the case that only 4 pieces of media isn't sufficient though. Hard disks have many moving parts, are are still susceptible to failure. If one of your backup drives dies you can probably get the hardware replaced by the manufacturer, but your data will still be gone. Granted, the same would happen if I had a tape cartridge die on me. But since I have a 28 tape rotation, I would lose much less data than I would if I had a 4-tape rotation.
Frequently when people are addressing backup/disaster recovery plans, they think that it's enough to simply have a backup. But if your media has gone bad (as you know), or you're not verifying your backup sets, or your backup solution is not resilient to a failure of some component then your backup becomes essentially useless. Nobody wants to wait until they need disaster recovery to find out that their plan didn't work as planned.
I wouldn't worry too much about the recovery time if a catastrophic server failure occurs. In your case you're talking about a software install versus software and hardware install, 10 minutes versus an hour. Either would be acceptable if a server goes to pieces.
Using your numbers, assuming you have 4 100 GB drives and back up 10 GB per night, you're looking at a little over a month's worth of data. Is that sufficient to meet your recovery and archival requirements? I don't know what industry you're in (or even what country), but in the US some types of companies are required to keep 2, 5, or even 7 years worth of data archived in a readable format.
The problem that you're having with your old solution may simply be that the tapes are worn out. Depending on the media used there are limits to the number of times that it can be overwritten (a lot of people miss that point).
You're right about having a rotation/subset of the full backup set on a single disk. That's the value of having multiple pieces of media no matter which format you choose.
I'd make the case that only 4 pieces of media isn't sufficient though. Hard disks have many moving parts, are are still susceptible to failure. If one of your backup drives dies you can probably get the hardware replaced by the manufacturer, but your data will still be gone. Granted, the same would happen if I had a tape cartridge die on me. But since I have a 28 tape rotation, I would lose much less data than I would if I had a 4-tape rotation.
Frequently when people are addressing backup/disaster recovery plans, they think that it's enough to simply have a backup. But if your media has gone bad (as you know), or you're not verifying your backup sets, or your backup solution is not resilient to a failure of some component then your backup becomes essentially useless. Nobody wants to wait until they need disaster recovery to find out that their plan didn't work as planned.
I wouldn't worry too much about the recovery time if a catastrophic server failure occurs. In your case you're talking about a software install versus software and hardware install, 10 minutes versus an hour. Either would be acceptable if a server goes to pieces.
Using your numbers, assuming you have 4 100 GB drives and back up 10 GB per night, you're looking at a little over a month's worth of data. Is that sufficient to meet your recovery and archival requirements? I don't know what industry you're in (or even what country), but in the US some types of companies are required to keep 2, 5, or even 7 years worth of data archived in a readable format.
- Status
Similar threads
- Locked
- Question
- Locked
- Question