Disaster Recovery

How do you do your back-ups?


In the past, tape was the solution. The problem is that this still tends to be the answer. Either backup admins are complacent and don't want to change, or decision makers are unwilling to look at new ways to do backups.

With the ever increasing capacity of disk, it's not a matter of not having the space. The price for disks such as SATA, while not the fastest on the block, is affordable, so it's not a matter of price. Could it be the speeds? I don't see how, especially since disks are faster than tape for spooling data. Disk can also handle more streams (more data can be written at a given time in multiple paths or from multiple sources). In general, tape can handle a single stream of data. (Backup software is changing this; just wait for the time you have to recover from the tape...). Maybe it's just a matter of understanding a disk-based backup solution. In that case, I'll go over the basics.

A standard tape backup process involves running a tape drive or library from a backup server. This server's role is to connect with the other servers on the network and pull their data across the network and place it onto the tape. Simple enough. Doing a backup to disk is just as simple. Most of the latest backup software can handle writing to disk as an alternative to tape. Change the process a little and have the backup server write to disk instead of the tape. The backup software can write faster to the disk since it's continuously moving, unlike tape drives that stop, rewind, and forward to place data onto the tape. Since writing to disk is faster, the backup jobs should complete sooner, meeting the ever shrinking back-up window. An even better benefit is that once the data is on disk, you have the rest of the day to get the data to tape for a second copy or for offsite (long-term) storage.

Disk level backups have another benefit: better restores! The real purpose of backups is not to just have the data, it's the ability to restore the data. For anyone who has had to restore files, this process can be long and painful. By keeping backups on disk (online) restores can happen much faster, since the restores are disk to disk. There's no need to call back tapes or to load and catalog them. A second benefit is that you can actually permit the users the ability to do the restores for you! (This is based on the backup software features.) While these are simple restores, imagine rebuilding a server that crashed. Assuming that tapes and backups ran correctly, this would still involve getting new hardware, reloading the OS, the backup agent, loading tapes, and restoring the data to the new server. With disk backups, there are no tapes to load and the restore is much faster. Should the backup process fail or the data corrupt, you'll know sooner and can adjust accordingly.

79 comments
tom
tom

We use Xatacom for online backups and wow! there customer support and software rocks! I would recommend them!

shuhrn
shuhrn

Agreed on getting off the tapes. I'm finally been allowed to do it as well. I'm actually moving all my data to a third party location that will do all of the upkeep, backup and maintenance, freeing up some of my time. All locations will be accessing this data from the third party, which will allow for faster transmissions then having everyone access my small little box in the back of the building. However, I'm also going to have an on site backup. By doing so, I should be able to run large queries much faster without interfering with others that might also be accessing the same databases. In cases that the third party goes down, we'll have it setup so it attempts their secondary, then my backup on site. Hopefully without any interruption to our different locations.

dogknees
dogknees

Our current strategy is to use a traditional grandfather-father-son system where you have a set of daily tapes, a set of weekly tapes(done on the Friday) and a set of monthly tapes(done on the last Friday of the month). Daily and weekly ones are recycled, but monthly ones are permanent and are never re-used. This has it's own issues as I describe below. These are all kept offsite except for the most recent ones which are retained onsite for 2 days. This catchs most of the restores we need to do without having to get tapes back from storage. We are currently still using direct tape backup but will be moving to a staged disk to disk to tape strategy soon. There are a few issues people should keep in mind regardless of the methods or media type used. Longevity. In a professional services environment, it may be a legal requirement to retain well over 10 years of data. This raises all sorts of issues with tape types and durability, CDs going "flaky", disks failing and so on. One way to approach it is to keep all the data, including the historic stuff online at all times. It gets moved to new hardware as servers are replaced, and is pretty much guaranteed to be accessible. Given the endless reduction of disk cost and increase in capacity and amount of data, the old stuff only takes a minimal proportion of the new drives. And, it's still backed up offsite. The time required to do a backup run also grows, and is the primary driver for moving to either disk-disk or continous backup where changes are backed up as they occur rather than at night. Disaster Recovery. This pretty much requires some form of offline storage. If you run a medium sized business, you really need to be able to get at your data within no more than a week in the event of a "complete loss" disaster. While things like billing may be able to wait, if you can't service your clients in a financial services environment, they will at least go elsewhere, and you may have no business to recover! Or they will start losing money or incurring fines from regulators. Even something like a small architectural firm wouldn't be able to continue managing the contruction of their clients buildings which could have huge flow-on costs. There are lots of challenges in building and running a backup strategy and new technology is always giving us new options. Regards

siddh_patel
siddh_patel

What will happen to the file being used while backup is Going ON. When you backup on this disk process will stop if file is being used or manually it need to be skipped.

teebes2004
teebes2004

At a previous job, I did tech support for a law firm (I am not a lawyer). I recall a vendor claiming that backups from tape were acceptable in court as evidence, as opposed to backups on hard-drive because backups to hard drives could be modified after the fact.

AstroCreep
AstroCreep

Currently, I work at a company with one main location (~150 desktops, 20 servers) and 30 off-site offices (average 5 systems each). I breifly discussed both of the below solutions in other posts, but here they are in full: For the off-site locations we are migrating them from crappy, old Travan drives/tapes to using a removable USB enclosure with alternating drives. We have them run NTbackup, then a script kicks off later that uses the command-line version of 7zip to compress the backup file and move it off to the backup drive. We also added logic to the script that will create a new directory everyday (ingeniously named after that day) to drop the compressed backup fil into, and it checks the drive for enough freespace and will delete the oldest directory. At our corporate location we're looking into a new solution. A year ago our old DLT loader died and it was going to cost 8 large to replace, plus our Veritas renewal was coming up. Our local service depot had a new option: a managed 'DPU' from Unitrends (link). Thing is we needed something RIGHT THEN, and my old manager thought the price was right (about $1300/month) - but it's not worth it, in my opinion. Yeah, the unit is nice and it seems to do the job well, but it's Linux-based (which I'm not at all familiar with) and it's managed by our service depot, so our 'control' over it is not as full as I'd like. So we're looking to build our own! What I'm thinking is we can do make a server with a RAID-5 with a few big SATA-II drives and have a couple of them as removable (for the 'off-site' factor). Here's what I'm thinking: -have each server run ntbackup to this backup/file server -during the day, have this server run a compression/encription function, then drop the backup files into a daily-created folder (similiar to what I described above for our remote sites), then delete the original backup files -have the server copy the most recent backup to the removable disk, which would be swapped out on a daily basis Anyone have any thoughts? Comments? Suggestions? I think it might work!

lafeyette_management
lafeyette_management

Level 1: All critical data is automatically archived nightly by the archive server, Liberty. Liberty has a 1/2 TB RAID 5 array and employs RSYNC and a special set of shell-script programs I wrote to perform a nightly full backup and then to make a 'differential' backup that retains the differences between the current full backup and the previous night's backup. A history of these 'differentials' is retained going back a user -selectable number of days (My machine's disk-capacity currently is good for about two to three weeks. This enables me to have a fresh full backup (the most recent one) as well as a 'log history' of changes between that full and the previous ones, so that I can roll back to any date in time simply by restoring one of those 'changes' files, or rebuild something to a specific date by restoring the full and then applying the changes from the appropriate date. Additionally, since this is an RSYNC-based backup, it is file-based, which means that I am able to restore any specific file or directory (or combination thereof) from any point in time for which I have a backup using simple operating-system level copy or RSYNC commands. Currently, this system backs up primarily Linux/UNIX hosts, but there are some Windows machines that are also backed up using this method. Eventually, I will be upgrading to a much larger capacity server and, at the same time, reducing the number of machines needing physical backup. I will continue to use the Snaptrol Suite (the shell-scripts that use RSYNC to perform the backups) as it has proven to be quite automatic and trouble-free, requiring little if any maintenance. It just runs.... Level 2: Critical data can be synchronized (even scheduled) across the network by using a custom shell-program I wrote called "parasync" (or its older, sequentially-based brother, 'multisync'). These two scripts simply use RSYNC to synchronize specific directories across machines whose names are provided as command-line arguments. Thus, for example, "parasync liberty compton ombostrsx102 ombostrsx103 octopussy lafeyette idaho-lan" would result in a few prompts asking which directory to syncrhonize on the host machine (the one I am running from), and the directory to which to synchronize to for the target machines (it is the same directory for all of them--one is not asked for a directory for each one). If the directory is missing from the target machine, it will be created. Finally, the program offers a choice between ssh and rsh for the means to communicate to the other hosts, and that's it. I often use this to synchronize home directories or data directories from one machine to multiple others. The primary difference between parasync and multisync is that the former runs in parallel and, therefore, cannot interact properly with prompts for username and password, so before using it, it is necessary to ensure that one can ssh (or rsh) to the target machines without being prompted for name and password. Multisync, running sequentially through the list of target machines is slower, but it can deal with prompts for name and password for each target host. OFF-SITE STORAGE. Laptops are among the target machines for these backups (as well as sometimes being hosts), and, as I almost always have one or another laptop with me (I know, I know. My wife says I have too many laptops. But I just don't understand why she thinks four or five is 'too many'....), that laptop's copy of the data is off-site. In addition, I generally copy data I care about to one or more machines at work (usually using the laptop to convey it) as well as using the reverse method to store critical work data (such as master copies of the VMs and other tools I create for my job.) Virtual Machines: I don't normally back up operating systems as these are fairly easy to reconstruct--my standard servers are as generic as possible, and, in most cases, backing up the operating system is pointless, because one must first reconstruct the operating system in order to get to the point of installing the software (or performing the copies) necessary to restore the operating system. Thus, in most cases, the task is 'recursively redundant'. However, there are cases where a heavily customized operating system installation is needed, in particular, for example, for VPN connections to my place of employment or for use as a golden master image for the virtual machines (over 100 and counting) that I maintain/create for my employer. In these special cases, I create virtual machine "appliances" and then keep multiple copies of these in multiple locations (home, work, laptops). This way, if a VM-based VPN client is lost or damaged (e.g. by a virus, ID-10-T error, or whatever, it is a matter of a few minutes to restore the VM and have a brand-new, perfectly workable VPN client in service. I am expecting, in future, to rely more and more heavily on sophisticated, highly customized VMs running on very generic, high-speed VM Hosting Servers. Thus, the backup techniques detailed above become applicable not just to files and directories, but to entire (virtualized) servers and workstations.

mhunter392
mhunter392

I am really frustrated by the lack of understanding shown in many of the comments about this critical business issue. It shows lack of mentorship and lack of IT/Admin education or research. Blogging for knowledge is only a speck of what you need to do. Hey, daily backups are done ??? ... DAILY ! This is the same regardless of the OS: MS, Novell, Apple, Unix, Linux, AS/400, mainframe, etc...... Usually a non-peak weekend night is a Full backup, with either Incremental or Differential backups done the rest of the week. Again, business operations have to be considered, so that they do not interfere. Then you have to do test restores, to verify that your backup operations work and the tapes are good. Tapes shoould be used for no more than 1 to 2 years. Tape Drives models should be selected appropriate to how many servers and how much is being backed up. Or do you do a SAN solution ? Backing up each server seperately as a stand alone is waste of assets (hardware, software, money, time, effort). Do centralized backups to a single Tape Library or similiar device if you have multiple servers. USBs ??? .... this is not your home network. These are only valid in the business world for a quick solution, not as long term/permanent solution. Still, I have only touched the basics. Get some education people - from a mentor, a local user's group, tech school, ... something from somewhere!!!! Based on the majority of the comments I read, it is scary how poorly some businesses are protected. If you have a restore problem and can not get back critical data, you should be fired. Protecting a business is serious stuff. Again, get out of the "home-network" or a "1 to 2 server" business mind frame and learn what must be done to protect the business. IT for IT sake is crap. IT is there for supporting business needs, based on how the business operates. My "Dennis Miller rant" is concluded.

junkdomain
junkdomain

i want to ask what should be best schedual for daily backup i mean which types of backup process should use.

Grimshiire
Grimshiire

Maybe management is to freakn cheap to invest in a decent backup solution than to get the freebee that comes with the Backup Hardware. I guess I should be thankful that I am not using Mag Tapes!?! Last IT Rites

mhunter392
mhunter392

Are you kidding? This option is fine for a home network, but NOT as a business solution. True Disaster Recovery is not only about doing fast backups & restores. You MUST rotate backups off-site. You MUST be able to recover a business data center at a separate facility. SAN solutions are a viable option or using tapes, not what the author suggests. I only touch on a few key of the many issues that must be addressed to protect a business.

Drew McClure
Drew McClure

We have 30 servers we backup daily with 800GB of data. We save money on agents by using Automate to run scripts to copy the data from these servers to a 4 TB drive on the backup server then backup that drive. So, we essentially have online, nearline, and offline backups. We are using a HP storageworks robotic library which supports 30 800GB DLT tapes.

robert.lutinski
robert.lutinski

Well... Yes. In this day and age, there seems to be little reason to invest in multi-thousand dollar tape solutions, when that same money can be spend on some form of drive solution. The only advantage a tape system has is that it is portable and great for offsite recovery. this becomes a moot point when the previous night's tapes aren't brought offsite. I think a better question would be is is better to create a cheap raid centralized backup server solution, or several smaller drives attached locally for backup purposes?

TexasJetter
TexasJetter

We use removable drives for offsite backup. Our scenario goes like this - Each night all data (which includes NT backup files of key servers) is copied to a removable hard disk (currently about 320Gb). We have three of these, one in the machine, one with me, and one at home. - Offsite storage solved. Each night the same data is copied to a "non-removable" JBOD store that has the last Tue, Wed, Thru, Fri. - Quick restore solved. On Monday all data is copied to a removable hard disk that is kept permanently. - Historical archived solved (we don't have to comply with SOX) We use RoboCopy with a mirror flag so that the nightly data does not creep up as users remove files. By using RoboCopy instead of any backup software we have immediate restore, just browse to the data and copy. We do not have to wait for the sequential read from a tape, we can restore multiple different folders simultaneously. Backup times have decreased as well. We were getting up to 12 hours for tape, using disk to disk copy brought it down to 6-7 hours. Cost wise hard disk come out cheaper. Initial cost is less, we were looking at several thousand dollars to increase our last Tape drive capacity vs. an existing machine with some removable trays. Also our standard has been to get two tape drives so that if one fails we can still read our tapes (makes a nice $2000 boat anchor). Ongoing cost is less as well - we were paying $60-70 for each tape, I just ordered some 320Gb for under $50. Storage is a bit of an issue - one hard disk is equal to two tapes, but storing 10+ years of tapes is difficult anyway. I would judge durability of a hard disk better than a tape (ok I agree with the previous post about dropping it) - if there was a fire which would you think would survive better. Also long term recovery for a hard disk I feel is better. I have three "old" different tape drives kept just to read our tape archive. Even with the hardware getting the software to read the tapes would be problematic. Hard disks do not have that issue, as long as the OS can read NTFS we are set. So to us it came down to less cost, more reliable, and easier recovery.

wdeckert
wdeckert

For Windows XP machines at home or work, I use XACT Copy from Duocor (http://www.duocor.com/). This system utilizes a second hard drive of equal or larger size to duplicate the entire hard drive including the OS. First drive fails, just unplug the failed drive and reboot. Second drive comes up immediately. I have utilized XACT Copy for years. It works. Want off site storage? Remove the second drive after a backup and store it off site. Install another drive and do another backup. It is fast and recovery is instantaneous. They are working on a version for Vista which will come out soon. The beauty is that it can work like Raid 0, but I prefer to manually schedule my backups after I do a virus scan and determine that my computers are running perfectly. Check out the site for more information.

edwardwstanley
edwardwstanley

with the risk of mediums being more fragile and the importance of backing up IT infrastructure waning, I've developed a new method of demonstrating the importance and size of data that makes a business happen. Backing up to "hardcopy". All Server and clients will now be backed up to paper. Were doing binary dumps where they will then be printed off and stored in tippet richardson boxes. For parity checks the back page will be the exact opposite page, eg page two will have page eof-1 page on the other side. A few dozen Highspeed Adf scanners with OCR binary input will be used to restore from. Current estimated warehousing space is only another warehouse = to the size of the current warehouse, so it makes sense that we put another floor on top of this building as not to disrupt current operations. Pt 2. Further discussion has shown that although our implimentation of the hard copy back ups have been quite successful, the issue of snapshotting the state of live data is an issue. A second technique is now chosen as per ITIL v3, where we have procured 400 4 GIG Memory cards and have created a RAID 5 out of them. It takes a good deal of time to build the raid, we estimate its completion sometime mid 2008, but were confident we will find it effective. Edward

Don Ticulate
Don Ticulate

Well... Local backups to disk AND backups to tape every night. The tapes get rotated off site but not the disks. Major Servers are also imaged to removable disks once per week or after a major change - just in case. The critical functions of the system have been replicated on a couple of VMWare servers that are off site. Although a bit slow in performance (due to cost), these servers have the ability to read both kinds of tapes used & the disk backups if needed. We even have VMWare desktop images should the need arise (with VPN software loaded).

rswain
rswain

Additional considerations for Tape. Small size and weight. Allow easy offsite storage which means some handling to the site. You drop them they will still work next time. Try that with a removable disk. Backup capacity is unlimited. Keep buying more tapes compared to limit on size of disk storage.

athome
athome

Hi, Weve written our own solution - AcBackup. This does indeed use extra hard drives but in a different machine. Heres how we work it - and I think its a good methodology. We have a spare computer connected to the network. Everyday at a given time, the AcBackup application starts itself up. The first thing it does is to clear space on the spare computer (call this the AcBackup comp), It then proceeds to take a complete copy of all files requiring backup = this is definable. The AcBackup computer actually keeps a full 7 days cycle - therefore on that machine should you need to restore anything you simply look within the Monday, Tuesday, Wednesday, etc folders. Now once the local copy has been completed, it starts all over again but this time to our secure off site servers. However, this time round it simply uploads to the off site server only files that have changed. Reports are produced every step of the way, logs are kept. And to round it all off - the nice little AcBackup computer sends us a text message informing us of how many files or errors where encountered. These are just some of the txt's from last night. 09/08/2007 - 21:55:01 - Finlay - Usb - Local (25112,1) 09/08/3007 - 20:43:19 - AcFinlay - Local Copy (25112,1) FTP Upload (29,1) 09/08/2007 - 19:38:04 - AcAj - Local Copy (31,1) FTP Uploade (10,1) 09/08/2007 - 19:02:21 - AcRbs - Local Copy (16130,1) FTP Upload (18,1) This systemn I believe offers everything, you can instantly restore onsite, instantly replace your server if need be and youve even got an off site copy of all your data on secure servers should the entire building burn to the ground - your insurance company will love you for that ! We do indeed provide off site backups for external companies as well as our own inhouse systems, we also develop applications, websites, etc. www.hestor.com - 0800 987 8140 or 07811 683 922 www.anywheresupport.co.uk www.autothumbs.uksellingpoint.co.uk Hope this offers a little insight - Dave. Remeber computers are here to make your life easier, so lets use them in that manor !

jgaskell
jgaskell

RAID is not a backup solution, it is a fault tolerance solution. Please tell me you understand the difference.

damone
damone

It doesn't matter if the source is disk or tape, either can be modified. The only solution is a device that cannot delete or modify the original data, or allow even an administrator to modify. These devices are few and far between and usually are expensive.

lawlormj
lawlormj

I thankfully do not speak from experience but, supporting a health sector client, I've had to consider this issue as well. Unless one is caught up in a full, corporate investigation then I believe that a court-ordered request has to be specific and that it can demand only material that relates, for example, to "Smith vs Jones" or "Patent No. 17" or "Correspondence conducted with Mr X between 2004 and 2006". To provide an entire media volume that contains information other than what is required may expose the ordered party to a host of other transgressions such as a breach of confidentiality. This means that the ordered party is on its honour to selectively extract the pertinent information from its backup media and to provide only what is asked for, possibly in another media format altogether. Is one media format any more inviolable or "permanent" than any other? I doubt it. Honour, it seems, still plays a large part in the legal issue. I'd appreciate hearing more on this matter.

forwejohn
forwejohn

This topic discussion has been a great source of information. However, my company is wanting to get away from manually copying weekly Full backups to USB drives in order to provide offsite storage. Total Full backups of all servers exceeds 160GB. I'm look for some guidance one alternative options to get these Full backups offsite without going back to tape. I've looked into setting up an FTP server at an alternate location, but I'm concerned about being able to copy the amount of data in a timely manner. Any advice?

damone
damone

One option to consider is a snapshot of your VMs. This would allow you to rollback to a good working version of your VM, instead of having copies all over the place.

AstroCreep
AstroCreep

Maybe it's only me, but I'm having a tough time fully seeing where your rant is going - simply decrying the 'USB' solution for a centralized data repository, or holding high the banner of the tape drive? Reason I ask is Unitrends has made a big name for themselves as of late with their "Data Protection Units"; a colleague of mine has been to quite a few MS seminars lately and many people are singing the praises of this thing. We currently are using one, but I don't like it much simply because of the cost we pay for it's support (like $16,000 per year); only reason we went with it is because: a: our old DLT loader died and it would cost as much to replace as it would have to repair (a good $8g) b: our Veritas stuff was coming up for renewal, another big expense c: the unit is managed by our support depot, so it removes the hassle of configs & checking. Nice thing, if you have the cash for it (not tape-based, either); were kinda cash-strapped, so we're looking to make our own solution, and I'm thinking about doing something similar with a bunch of SATA drives in a RAID-5 array with a couple removable drives to be used for the off-site/archival portion of the backup solution.

zyphlar
zyphlar

Yes, daily inc/diff and weekly full backups should definitely be done. Centralizing backups to a SAN or NAS reduces redundancy. In addition to our NAS, we use those hated external USB drives for our weekly offsite backups and monthly historical archives. I've found this to be much more efficient than tape, because they're harder to damage, faster to write/catalog/restore, and it's just file copying-- no jobs to schedule or sets to overwrite.

jgaskell
jgaskell

Well put. There is a difference between a "computer guy" and an IT professional, but there are plenty of the former who think they are the latter.

zyphlar
zyphlar

Incremental or differential is my preferred since it has less overhead than full or copy, but covers a more complete range than daily. By microsoft's definition daily is from midnight to midnight, so I would worry about capturing data modified after the previous night's backup but before midnight. You might as well use incremental and have that base covered.

jc@dshs
jc@dshs

With the cost of USB drives "so cheap" why not just substitute 10 USB drives for the old 10 DAT tape cycle. 4 USB drives incremented daily, full backup on Fridays, overwrite the Mon to Thurs each week, keep your 4 Fridays until the end of the month, full backup on last Friday that you keep until end of next month whilst recycling the other 3 Friday drives each month,...

zyphlar
zyphlar

A tape drive, Veritas, and woefully-small tapes will run many thousands of dollars. My solution, using NTBackup and 500gb external hard drives, costs only $200 per "tape" and comes with an "Exchange connector." Cheaper, and in my opinion, creates less headaches.

damone
damone

I absolutely agree that there are many points to address for business protection. Backups and restores are more of a potential that we all will face. Disaster recovery is a business protection process that must be looked at and address based on business needs, not just data recovery.

zyphlar
zyphlar

How much did that library run you, and how many hours a week does your backup management require? Out of curiosity what backup program do you use and do you recommend it?

ken_stephens
ken_stephens

I use Acronis to a backup server with a Firewire attached external Maxtor 750GB drive. The files on the Maxtor are copied to 2 Maxtor Drives; 1 goes to a fire safe onsite, the other is sent offsite once a week.

lawlormj
lawlormj

Described elsewhere here is a predicament that I encountered with HDDs being kept offline for long-term storage: of about 25 drives only 5 were accessible - the majority powered up but would not spin while the data on others was inaccessible for other mysterious reasons. The drives were all 80gb, from a variety of respectable manufacturers and they were stored properly. Some may have been idle for as long as two years. New, unused drives seem to age well but it may be that once a drive has been in regular use and then retired its mechanics become unreliable.

triniweb
triniweb

We have been exploring the option of disk based backup solutions to solve the primary problem of speed of tape storage. There are additional cost to making Disk storage compliant with mentioned DRP and BCP typically would require an offsite solution as well as a reduced (spread) risk pattern for content of backup ie no eggs in one basket solution such as a single drive for all our backup data. Disks are considered prone to failure vs tape, and using their full capacity to store multiple sets of data is even more risky. Hence RAID, remote server NAS or SAN at DR site, or multiple RAID storage packs to be transported physically to DR site. with this in mind, our solution options were: 1. Portable NAS (RAID) -- 2. Remote DR site with SAN using VPN WAN link to hot or cold site -- cost of additional site (some offsite storage facilities/vendors offer this option now eg our cable/internet company has bunker sites and rackspace and offer site to site fiber data links or vpn over cable), hardware, and link 3. Co-sharing with another organization NAS/SAN over wan (one NAS/SAN each at 2 sites) -- only cost is data link -- I would love the option to have drives mobile but SATA/IDE and even SCSI drives have not proven very robust after moving them around for a couple of weeks between sites plus I would hate to have a bad sector and the lack of separate media for previous day/month make my life a living hell. RAID helps but makes for a bulky package however, transporting a drive still seems to be a secondary or temporary measure so it has been included with the preference for option 2 or 3 as a supplementary option for if we have problems with remote link to site (which by the way should have redundant paths *this is ofter offered as well). The maintenance costs for additional site may increase in all options just for warranty and administrative charges attached to monitoring QoS - link, storage solution, etc. Tapes do appear mush simpler and cost effective in the long term but that should change with reducing bandwidth costs and more data link enabled DR centers, more options for cosharing with increasing DR aware IT centric companies, cheaper and more streamlined remote storage solutions. I really would love revisit this when the price of Solid State Disks decreases and their capacity increases... they definitely will make me think twice.. abut the fragility of disk storage vs tape.

MWRMWR
MWRMWR

well there's not even a mention of 2003 server and the news items haven't been updated in 6 years. Maybe the group has been swallowed up and the originators are living in luxury somewhere ? If this product is really active, then I would be really interested but I'm afraid I just see alarm lights flashing. Also, how does it score over a mirrored raid system where you just buy a new disk from time to time and re-build the mirror ?

kraken_
kraken_

How strange the same ideas does pop up in differerent time and places :-) We also developed a -simple- application that was specially tailored for our needs, using disk backups. We too used a "7 day retention on disk" scheme for ease of recovery. The main difference is that back then, offsite backup using the network was not an option. We did keep the data on the disks, but a tape backup was connected to the server. When the backup was complete, the day's worth of data was transfered to tape quickly, without the "shoe-shine" problem (when the data comes in too slowly, the tapes stops, rewinds a few inches, then starts forward again.) since the data was available locally. The tapes were then sent to a vault (offsite storage). The users could do some restore jobs by themselves (the data from the user's PCs was shared on the backup server, but only to the pc's user). Both users and sysadmins were happy with this solution.

damone
damone

A lot of things depend on honor. In my experiences, business professionals view IT guys as "one who can fix it" or "can make stuff diappear." The problem is a lot of us have morals and ethics. Sure I know how to beat the system, but it is not right and there is pressure to do what the company asks, because if you don't, there are consequences. That being said, you have to look to technology to protect us. This means devices such as DVD-ROM and WORM (Write Once Read Many) devices. These devices can write, but usually never delete data. The data is intact in it's original format. If called up and modified, a new file is recorded, while the original is still in tact. So the technology exists, but how do you get around that? Well, DVDs can be destroyed (grab with two hands and SNAP). Disks can be erased. The solution has to be within the devices to protect from these manual manipulations. So far these protects are limited. We have to hope for a solution in the future, and soon.

mhunter392
mhunter392

What you do fits the business need. There is no cookie-cutter solution, as you apparently know. My "rant" was only towards the "home-small" network guys. Word has to get out there that IT is there to tailor itself and support the business needs. This topic is part of a huge function in what it truly encompasses, not a "sound bite". Thanks.

mhunter392
mhunter392

Thank you. I ranted enough. It is upon us to make sure that passing on the knowledge and skills to the younger crowd is done. You probably have experienced mistakes as I did when I was a young pup in the field. Experience & training & mentorship help business recover from mistakes and/or disasters. Problem I see is, our peers are not investing the time to mentor their subordinates or others in a forum such as this. Enjoy !

Drew McClure
Drew McClure

This reply is late but here's the info you requested. The library was quite expensive, around $4,000 just for the enclosure and about $130/tape connected via SCSI to the backup server. The backup server is running Backup Exec 11d and it's damn good (especially since the release of SP1). In terms of the management it took awhile to get all the scripts set up on all the servers but management is limited now. When new servers are created we just copy the scripts over.

Jim_MacLachlan
Jim_MacLachlan

I'm revamping my new, very old network. All (3) new Windows 2003 servers with 200gb on one & 50 on each of the others. I'm looking at tape systems & they seem to run around $5,000 to start just for the hardware & $2000 for Veritas. Acronis is half that cost & 500gb USB drives would run me just over $4000 for 25 of them. From the discussion above, I can't see a downside in my situation, but I'd sure like to talk to a couple of people who actually use it. I'm afraid of hidden 'gottchas'. Selling tape & Veritas to the suits is easy -it's standard. Selling anything else requires a lot of due diligence. I plan on getting a trial of Acronis & using my 500gb drive to test somewhat, but I can't create a separate Exchange server & restore email - no time & materials.

bobwh
bobwh

It's been my experience that in a small business without an internal IT department, a backup process that isn't stone simple will fail. The only thing that has consistently worked is to backup all servers to a single media every night, on at least a 2-week cycle, and store the media off-site or in a fireproof safe. Whether the media is disk or tape depends on time and expense. I know it's inefficient, but if they have to keep track of which days to use which media, or try to figure out which incremental media to restore from, fuggetaboutit. Hot-plug RAID drives are a virtual necessity. Check the status lights when you change the tape. Make sure you have the ability to a)boot from a backup tape or b)boot from a CD, to do a full server restore. Recently, I've put some clients on the iDrive online backup service, and so far so good.

jbuckner
jbuckner

Send it off-site on line (to a firm like Intronis), + use RAID 1 (inexpensive) and use an external HD on-site with Symantec. Rotate or retire External HD's as required. Call it a night (or a year).

zyphlar
zyphlar

Keep in mind this is a small business, about 100 users. We won't need to use the drives unless the backup server is unavailable or we need to access archived history. In the dr scenario where the server room is destroyed, a usb drive is also simpler to access than a tape.

Genera-nation
Genera-nation

but I would worry about only having a disk based backup. I currently have both disk & tape!

zyphlar
zyphlar

That our "tape drive" cost $0, and cataloging/indexing/retensioning is near instantaneous. plus the tapes can be used as a restore destination in a pinch. all those added features bring up the price a bit. Maybe I'm biased, but I don't like waiting for 20 minutes to see whether or not my data exists.

zyphlar
zyphlar

I don't trust image-based backup solutions, because if you run the backup on a live server you won't correctly back up any databases (Exchange, SQL, etc) running on the server. To correctly back up Exchange you either need to shut off the services and then image the files (offline), or use an Exchange connector that uses Exchange to backup itself (online). My backup solution, after years of Veritas and tape, is to use the Windows Backup utility and then copy the .bkf file to an external hard drive. So far the only "gottcha" has been the Exchange issue and the tendency for such backups to fail over a slow network link or during a server restart (but that's probably common to most other solutions.) I'm not an expert though, so if anyone has feedback please leave a comment.