September 1, 2009 at 3:00 pm
IceDread (9/1/2009)
GSquared (9/1/2009)
I were not after trying to bash WHS all thou it's not stable enough and not fast enough for me to even consider.
You seem to have an odd definition of "bash".
I would consider bashing if I talked about everything that WHS can do how it does it and then show why that is all very very bad. I didnt do that, I only talked about a few parts that is not good enough for me but I didnt try to claim that they were not good enough in general.
Don't mind me. I just get a kick out of irony. The statement I quoted is most certainly ironic.
Since I've never used WHS, nor any comparable service/product, I don't really have an opinion on any of them one way or the other. Hypocritically, I do have opinions on people I don't know personally. I laugh at myself, too.
- Gus "GSquared", RSVP, OODA, MAP, NMVP, FAQ, SAT, SQL, DNA, RNA, UOI, IOU, AM, PM, AD, BC, BCE, USA, UN, CF, ROFL, LOL, ETC
Property of The Thread
"Nobody knows the age of the human race, but everyone agrees it's old enough to know better." - Anon
September 1, 2009 at 3:10 pm
First, let me say that I'm no WHS fanboy; I simply find it the best tool for what I need (a home server). I'm rather agnostic overall; I use and appreciate a variety of Linuxen, Solaris, and many flavors of Windows on a daily basis: all have strengths and all have weaknesses. As that oft-repeated net quote goes, "Why can't we all just get along?" (Rodney King's legacy!)....
BUT...
IceDread (9/1/2009)
My summery of why I don’t want WHS is:1.What I know about it’s storage it’s not safe enough for me, raid1 is not something I like to trust.
Well, first, as has been amply stated in this thread, WHS 'raid' isn't really Raid-1; if you have 3 drives on WHS and loose 2 of them, you probably only loose 1/3 of your data. With a Raid-5 system, you would loose it all. With ZFS you loose zero, but the WHS system gave you 50% more useable space than the ZFS system, so that's not a comparison of equals.
But all of the above is the theory; let's think about reality: this is a home we're talking about, so recovery time constraints aren't exactly the same as a mission-critical business. (Leaving spousal issues aside!) So you recover any lost files from your offsite backup, which on a home i-net link takes some time. Big deal. In my case, I might loose some recorded tennis TV and/or ripped media: I don't send those offsite (too expensive). Oh well! But the key point is that the likelihood of simultaneous drive failures versus the pain of recovery or loss is pretty well balanced in WHS' favor in my mind -- for the home evironment! Add in the ease of management - and an alert on all computers in the house that a drive went south! - and WHS wins hands down.
2.Reads and writes speed you could check up on yourself but know that 200mbyte/s both ways is far from uncommon these days for ZFS servers. This is way faster than WHS btw from what I’ve seen in tests, all thou WHS might be fast enough for some people, depends on which features you make use of. For ZFS you will wanna make use of http://en.wikipedia.org/wiki/Link_aggregation untill you can get your hands on 10gbit/s cards because 200+mbyte/s reads is more than your 1gbit Ethernet card can handle.
Um, this is a house we're talking about right? A single-family residence? My humble abode is pretty much media central with all the ripping and DVR'ing and hi-res photography I do, yet I don't have a bandwidth problem. And I'm running a rather slow server: an old Dell desktop, ~1GHZ single core (remember those?!), and 16 drives on 4 SATA ports via multipliers, with one PCI32 bus. I do have two Gb links on the machine teamed, but that's the only tip of the hat to the speed witch. And lo, it's not unusual for this machine to be recording two high-def channels while streaming 3 MPEG-4 or MP3 sources and moving some files. And not a dropout in sight. Why pay for bandwidth that you won't use?
3.Backup, at home, really? Well I actually do that too but to a very small extent. I like backups on another location. I do not want to have the same file on both my laptop, my tower and my server. I want it only in my server, this I didn’t catch WHS doing. This means I will have more space available in total. There is no point having a file in several places except for a backup outside of your home. A raidz2 can handle enough disk crashes for me to sleep well, WHS this far does not comfort me.
There are several solutions for off-site backup with WHS; at least two are available as WHS 'add-ins' (integration into the management GUI). WHS has neither a substantial advantage in this area nor a disadvantage relative to any other OS.
Trust trust trust, I can’t trust raid1 to keep my data safe, not to mention it’s not space efficient, and WHS from the tests I’ve seen is slow, making use of only 1 drive to read from probably thus the slow speed.
Well, the speed issue is moot, and as the WHS isn't really raid-1, it doesn't follow your assumptions about raid-1. Importantly, 3 drives on a WHS system have useable space equivallent to 1+1/2 drives, but in the event of drive failure, the data loss is either zero (one drive fails, possibly also true if two fail but depends on probability), one third (two drives fail), or all (all fail). With my 16 drive WHS system (8TB full useable), one drive failure has no loss, with probability of data loss rising with additional simultaneous drive failures from a very low value until virtual certainty at around 12 drives failing, depending on how fully the 'array' space is used. (Analyzing the probabilities is actually very interesting.)
And this now brings me to the one thing that you have completely overlooked, and which MSFT took great pains to get right: in a home environment, storage needs are ever-expanding -- space, the final frontier! With WHS, add a new physical drive, click on the GUI button in WHS, and you're done: drive added to pool, storage balanced, more space available. Any size drive: same as the drive(s) in the system, bigger, smaller, huge, tiny; it doesn't matter. In my case, with the Addonics port multipliers, I can tack on a new multiplier (fan out of 5, depth almost unlimited) and drive(s) in minutes!
With your ZFS solution? Um don't look now, but there's no incremental growth capability in any ZFS implementation to date.... So instead of incremental growth, you re-buy a full suite of drives to hold your new target size, then install them all, then build and migrate, and decommission the old ZFS. Or hey, with all that work, might as well migrate to a new server platform to maintain the bandwidth in line with the storage volume. Wow, sounds just like work - just plan the capital allocation! But hey, you will have way more disk bandwidth than my lowly WHS, you will have a 'real' raid system, as your space needs grow your bandwidth capability will probably grow, and you will be fully immune to simultaneous drive failures. Awesome!
September 1, 2009 at 4:00 pm
Steve Jones - Author (2009/08/21)
Do you use RAID at home?...Our data is important to us, and we want to be sure it's backed up and protected. So far I have 6 USB drives with various data spread around the house, some of them duplicating others, but I've decided to move to the Home Server model and their Drive-safe technology.I'm curious what the rest of you think and what you use.
The difficulty of setup and configuration of RAID seems to have a rather wide margin of perceived difficulties within this discussion. I think if you're going to answer this question, you need to at least be aware of the skill set of the audience from whom you're expecting an answer. Not so?
September 1, 2009 at 5:10 pm
IceDread (9/1/2009)
IceDread (8/30/2009)
And I still can't understand why pick WHS, a system that has proven to have quite a few bugs, some file types not compatible etc. Not to mention hardware raid is slow and binds you to that hardware. All thou I'd probably do WHS if it supported ZFS but seeing how it does not have the latest, the best, no WHS for me.After that peeps gets flaming mad with me as you see if you read the thread.[/b]
SQAPro (8/31/2009)
IceDread (8/30/2009)
And I still can't understand why pick WHS, a system that has proven to have quite a few bugs, some file types not compatible etc.Care to provide a source or two for that alligation? otherwise it's just slander and hearsay. .
If you do not know about the file corruption bug then you should search for it because it’s common knowledge these days and if I provided a link it would just be claimed fabricated would it not? Anyway here is a link. http://www.anandtech.com/weblog/showpost.aspx?i=413
Fixed. Over a year ago. http://support.microsoft.com/kb/946676
Care to try for something that actually currently affects WHS?
I mean if you want to use old fixed issues that were present in the initial release as your basis to call something buggy and unreliable, I'm sure we could, by that same standard declare that ZFS is 'buggy and unreliable'
September 1, 2009 at 5:23 pm
Sir Slicendice (9/1/2009)
Well, first, as has been amply stated in this thread, WHS 'raid' isn't really Raid-1; if you have 3 drives on WHS and loose 2 of them, you probably only loose 1/3 of your data. With a Raid-5 system, you would loose it all. With ZFS you loose zero, but the WHS system gave you 50% more useable space than the ZFS system, so that's not a comparison of equals
You also are presuming in that scenario, simultanious drive failure, and no other backups for there to be data loss. Potentially in a non-simultanious situation, depending on the size of the drives vs the data stored there, I believe that (I'd have to check with someone at MS to be sure) whs could have re-mirrored the data, and you would even after the second drive failure be out zero data. A lot of that depends on how much excessive drive capacity you choose to afford. I don't see most home users going for a 'hot spare' drive.
And as you point out there's the consideration in a HOME environment And the cost of a multiple redundant drive system that the average user has no hope of setting up much less recoving from a failure mode vs how valuable the data is, and if it's potentially backed up offsite anyway either via DVD's at the office, an online backup service, or something like photobucket etc.
ZFS might be a great solution in the land of theory, but it fails to transition into the land of practicality for the average user.
September 1, 2009 at 5:35 pm
Most of the audience here are database professionals. I can assume they understand and can set up RAID systems.
I'm asking what they use at home, and if they want to mess with RAID.
September 1, 2009 at 6:49 pm
IceDread (9/1/2009)
IceDread (8/30/2009)
Which brings me to the same link above talking about the software raid in WHS, unless you download the Drive Extender info from ms homepage which I did. It looks like raid1, and that is not something I want to do. What I want is Raidz2 which I think requires min 4 or 5 disks thou I recommend doing 6 disks so you only lose 33% of the data to the raid and have 66% of the drives space to make use of. Out of these 6 disks, 2 can fail and the zfs still works.
6 disks... heh yeah there's a solution applicable to the average home user.
But still, to compare, if you had WHS with 6 drives, we would figure that each drive would have one copy of roughly 1/3 of the data you designated important enough to store redundantly. So if two drives were to fail simultaniously, then the system would potentially lose 1/3 of your data, presumng the two drives that went down held the exact same data. However the odds would be against that. roughly you'd have 1:5 odds that the second failed drive also held data that was on the first failed drive. Presuming WHS scattered stuff randomly among the drives your loss would be 1/15th of your data.
That's also presuming some event which causes two drives to fail at the exact same time, but doesn't take out the entire system in the process.
My summery of why I don’t want WHS is:
1.What I know about it’s storage it’s not safe enough for me, raid1 is not something I like to trust.
not even at home? with offsite backups as well? Please remember the entire context of this discussion is what people are doing AT HOME.. not for business or commercial needs, but at home. I get that you personally don't trust mirroring, but seriously do you think there are no situations for which it's an appropriate solution?
2.Reads and writes speed you could check up on yourself but know that 200mbyte/s both ways is far from uncommon these days for ZFS servers. This is way faster than WHS btw from what I’ve seen in tests, all thou WHS might be fast enough for some people, depends on which features you make use of. For ZFS you will wanna make use of http://en.wikipedia.org/wiki/Link_aggregation untill you can get your hands on 10gbit/s cards because 200+mbyte/s reads is more than your 1gbit Ethernet card can handle.
and therefore WAY beyond the capacity of the average home user who's typically running 100baseT or WiFi.. so completely wasted in a home environment.
I'll grant that's very cool to get high performance like that in a situation where it's warrented.. but again where's the value in that much speed at the server when we are talking about home users with single nic systems on a non-cutting edge network.
3.Backup, at home, really? Well I actually do that too but to a very small extent. I like backups on another location. I do not want to have the same file on both my laptop, my tower and my server. I want it only in my server, this I didn’t catch WHS doing.
showing once again that you really don't understand WHS. It's perfectly able to serve as a traditional file server (and print server). Furthermore when you create the shares you can determine if they will be stored with or without fault tolerance, so you can save disk space for stuff that's transitory or easily replaced and doesn't need to be stored mirrored.
Trust trust trust, I can’t trust raid1 to keep my data safe, not to mention it’s not space efficient, and WHS from the tests I’ve seen is slow, making use of only 1 drive to read from probably thus the slow speed.
we get you don't trust mirroring, heard you the first time. WHS isn't designed as a product for business where speed is an issue, it's designed for the home user. To that end, even a single SATA 1 drive is faster than a WiFi or 100baseT network, so unless someone has a gigabyte network at home, WHS ought to be more than capable enough of being 'faster than the network' and hence not the primary bottleneck
It sure looks like I insulted someone's religion because things like this is common to see in those cases.
It's not religion, and frankly I think that's a copout on your part because saying 'oh I offended their religion' allows you to ignore the real reasons for the turmoil.
My objections are not that you propose a linux or non windows based solution. My objections are:
1) disinformation regarding stability and reliability of WHS (as I've said previously), all based it would appear, on a bug fixed well over a year ago, and your opinion that drive mirroring of any sort is 'not reliable enough' despite the fact it works quite well for a large number of users and applications
2) criticizing a product that you clearly don't understand both in terms of features and/or capabilities, basic operation, and intended market.
3) promoting an inappropriate solution which is in almost all ways irrelevent to the discussion that was taking place.
I mean seriously, here we are talking about a backup system and/or redundant storage for the HOME envirohment, and do we think RAID of any sort is too complated for the home user, and you come in and start talking about something based on open solaris OS with a 6 disk drive array.
I mean it's a little bit like if this was a car forum and we were discussing appropriateness of Honda Civic vs WRX and Mazda3, and you come in and say 'those are all trash, I don't trust them, I'm going to build myself a Zonda R roadster' (cause the maker released the plans a while back. although that analogy gives a bit too much credit to Sun to compare them to the likes of Pagani, I think it serves to make my point.)
September 2, 2009 at 12:27 am
ssssssssssssssssssss
September 2, 2009 at 12:31 am
Sir Slicendice (9/1/2009)
IceDread (9/1/2009)
My summery of why I don’t want WHS is:1.What I know about it’s storage it’s not safe enough for me, raid1 is not something I like to trust.
Well, first, as has been amply stated in this thread, WHS 'raid' isn't really Raid-1; if you have 3 drives on WHS and loose 2 of them, you probably only loose 1/3 of your data. With a Raid-5 system, you would loose it all. With ZFS you loose zero, but the WHS system gave you 50% more useable space than the ZFS system, so that's not a comparison of equals.
Well, the speed issue is moot, and as the WHS isn't really raid-1, it doesn't follow your assumptions about raid-1. Importantly, 3 drives on a WHS system have useable space equivallent to 1+1/2 drives,
You are contradicting yourself here. Looks to me like you lose 50% of the space in WHS to the raid.
But all of the above is the theory; let's think about reality: this is a home we're talking about, so recovery time constraints aren't exactly the same as a mission-critical business. (Leaving spousal issues aside!) So you recover any lost files from your offsite backup, which on a home i-net link takes some time. Big deal. In my case, I might loose some recorded tennis TV and/or ripped media: I don't send those offsite (too expensive). Oh well! But the key point is that the likelihood of simultaneous drive failures versus the pain of recovery or loss is pretty well balanced in WHS' favor in my mind -- for the home evironment! Add in the ease of management - and an alert on all computers in the house that a drive went south! - and WHS wins hands down.
I disagree. But then I don’t like losing data which you seem to be ok with.
2.Reads and writes speed you could check up on yourself but know that 200mbyte/s both ways is far from uncommon these days for ZFS servers. This is way faster than WHS btw from what I’ve seen in tests, all thou WHS might be fast enough for some people, depends on which features you make use of. For ZFS you will wanna make use of http://en.wikipedia.org/wiki/Link_aggregation untill you can get your hands on 10gbit/s cards because 200+mbyte/s reads is more than your 1gbit Ethernet card can handle.
Um, this is a house we're talking about right? A single-family residence? My humble abode is pretty much media central with all the ripping and DVR'ing and hi-res photography I do, yet I don't have a bandwidth problem. And I'm running a rather slow server: an old Dell desktop, ~1GHZ single core (remember those?!), and 16 drives on 4 SATA ports via multipliers, with one PCI32 bus. I do have two Gb links on the machine teamed, but that's the only tip of the hat to the speed witch. And lo, it's not unusual for this machine to be recording two high-def channels while streaming 3 MPEG-4 or MP3 sources and moving some files. And not a dropout in sight. Why pay for bandwidth that you won't use?
Yeah a home, you know 1080p movies to tv’s and such things as well as playing music in another room and being able to use computers with high speed and no lag. But if you want to stay with the read/write speed of hdds then by all mean, you should never look at SSD drives because those are insanely fast, it takes like no time at all to load some stuff.
3.Backup, at home, really? Well I actually do that too but to a very small extent. I like backups on another location. I do not want to have the same file on both my laptop, my tower and my server. I want it only in my server, this I didn’t catch WHS doing. This means I will have more space available in total. There is no point having a file in several places except for a backup outside of your home. A raidz2 can handle enough disk crashes for me to sleep well, WHS this far does not comfort me.
There are several solutions for off-site backup with WHS; at least two are available as WHS 'add-ins' (integration into the management GUI). WHS has neither a substantial advantage in this area nor a disadvantage relative to any other OS.
But you do like to let the same files take up space on several computers…
Trust trust trust, I can’t trust raid1 to keep my data safe, not to mention it’s not space efficient, and WHS from the tests I’ve seen is slow, making use of only 1 drive to read from probably thus the slow speed.
Well, the speed issue is moot,
For you perhaps so you don’t have to look into things like SSD drives for your home computers.
and as the WHS isn't really raid-1, it doesn't follow your assumptions about raid-1. Importantly, 3 drives on a WHS system have useable space equivallent to 1+1/2 drives,
So you lose 50% of the space to the raid, which is not very effective, and you complained about raidz2 taking up space.
And this now brings me to the one thing that you have completely overlooked, and which MSFT took great pains to get right: in a home environment, storage needs are ever-expanding -- space, the final frontier! With WHS, add a new physical drive, click on the GUI button in WHS, and you're done: drive added to pool, storage balanced, more space available.
Yes this is a good thing for the WHS
With your ZFS solution? Um don't look now, but there's no incremental growth capability in any ZFS implementation to date....
Actually there is but nothing good what I’ve seen, so you have to plan ahead or change out the drives one at the time to a larger size until all drives are changed.
Wow, sounds just like work - just plan the capital allocation! But hey, you will have way more disk bandwidth than my lowly WHS, you will have a 'real' raid system,
Yeah I like speed and stability and I like not losing 50% of the space to the raid. Not to mention it’s unsure of how the whs handles the failure of disk crashes, it’s more unpredictable. But hey since you don’t like speed, you wont have to upgrade in a very long time. If the read and write speed of one disk is fast enough for you, good for you, people like me see the value in higher speed, probably why so many are happy about SSD.
September 2, 2009 at 8:14 am
Steve Jones - Editor (9/1/2009)
Most of the audience here are database professionals. I can assume they understand and can set up RAID systems.I'm asking what they use at home, and if they want to mess with RAID.
I'd have to say that the prior poster commenting that the perceived difficulty of RAID varying greatly is spot on. I would also note that some, but not all, database professionals, in their professional capacity, do not spent any significant amount of time on RAID setup beyond, in a subset of those cases, "This database should probably be RAID 10, and that should be RAID 5."
Personally, I would say that the order of complexity and robustness of home level RAID is roughly the inverse of the cost of that RAID.
Pure software RAID-5/6 with Linux is very cheap, and more complex than I personally want to spend time on.
Pure software RAID-1 with Linux I'm planning on playing with as a stopgap until I can budget a true RAID-6 replacement.
Intermediate firmware/software RAID I've never messed with at all.
Pure hardware RAID-1 isn't too expensive, and probably isn't too difficult.
Pure hardware RAID-5 I've been using at home since about 2002, and it's really quite easy to set up if you simply walk through the text based BIOS menus on the card. It also cost perhaps $400 for the controller in that year, plus the cost of a set of ATA drives.
Pure hardware RAID-6 is my replacement plan, and it should be of similar relative simplicity, and cost on the order of $800 for the controller, plus $100 for an LCD faceplate, plus the cost of a set of SATA drives. Advantages will include regular RAID consistency checks, which the prior card did not have available.
Speed is almost completely irrelevant - 1Gbps Ethernet is the bottleneck for sequential access, and for the rare random access requirement the budget constraints of magnetic SATA drives constrain me; a local RAM disk, Intel X25-E, or OCZ Vertex would fulfill any major random access requirement I had.
P.S. Most RAID that I've worked on does not _require_ drives to be the same size, it's merely constrained to the size of the smallest drive in the raidgroup. I once replaced a 75GB drive with a 120GB drive of another brand after a failure.
September 2, 2009 at 1:10 pm
I could set up RAID for home, but I chose not to. I chose the Windows Home Server solution because it would backup my systems automatically and perform differential backups. Also with the Windows Home Server solution I can do a complete restore to a blank hard drive if something was to happen to the hard drive in my system. Another reason for going with Home Server is that I can easily add hard drives to the system as I can purchase them in as the best price/capacity that I can get at the time. I like the additional features of providing remote access to all of my computers on my home network as well as provide network storage. Yes there were many options available from scripting, to trying to learn how to implement Linux to provide a free solution, but the time versus money equation ruled in favor of actually purchasing the Windows Home Server hardware offered by Hewlett Packard.
James E. Freedle II
September 3, 2009 at 11:02 am
SQAPro (9/1/2009)
Sir Slicendice (9/1/2009)
Well, first, as has been amply stated in this thread, WHS 'raid' isn't really Raid-1; if you have 3 drives on WHS and loose 2 of them, you probably only loose 1/3 of your data. With a Raid-5 system, you would loose it all. With ZFS you loose zero, but the WHS system gave you 50% more useable space than the ZFS system, so that's not a comparison of equalsYou also are presuming in that scenario, simultanious drive failure, and no other backups for there to be data loss. Potentially in a non-simultanious situation, depending on the size of the drives vs the data stored there, I believe that (I'd have to check with someone at MS to be sure) whs could have re-mirrored the data, and you would even after the second drive failure be out zero data. A lot of that depends on how much excessive drive capacity you choose to afford. I don't see most home users going for a 'hot spare' drive.
Absolutely correct - I was discussing the worst-case situation. And yes, WHS will re-balance immediately (subject to load, etc) on detection of drive failure; I've seen it in action. In a 3 drive situation, if you lose 1 drive, you lose 1/2 drive's worth of capacity (assuming all drives are the same size): as long as you have minimum 33% free space prior to the failure, you maintain full redundancy after the re-balance process is complete (which will require the amount of time it takes to write 1/3 the size of a drive, give or take). It's very cool....
True simultaneous drive failure is extremely unlikely in a conventional home server; with power shared with CPU and integrated controller, the only way you get to simultaneous failure is with simultaneous drive hardware failure which has a probability so low it's not worth considering (in a HOME environment!). In my situation - multiple controller cards, port multipliers, external disk racks - simultaneous failure can result from a number of causes (drive rack power supply, cable, port multiplier, controller, etc), so it was something I thought about. But with this sort of failure situation (bringing the drives back online), I was able to recover with no data loss thanks to the individual drives being standalone NTFS volumes (I used SyncToy to merge the files back into the main pool). But I think all this discussion, both the failure mode and my recovery approach, is beyond a typical home installation....
...
ZFS might be a great solution in the land of theory, but it fails to transition into the land of practicality for the average user.
Honestly, I've looked at ZFS for work, too, and while I'm thinking about having a test/demo server set up, I'm still not fully convinced. The issue for me is that while I gain some space advantage versus Raid-1 or 10, my infrastructure team has a significant learning curve to climb and I'm not sure that all the tools are in place or even available to manage the ZFS environment as well as we manage our 1/10 environment. And while more space is nice, in the overall calculation, space is still really cheap relative to labor and productivity and especially business risk, even when I factor in rack space, cooling, power, controllers, etc. (We don't do raid-5 because of problems with simultaneous drive failure [failure during re-build process] and recovery time constraints.)
September 3, 2009 at 11:31 am
Sir Slicendice (9/3/2009)
True simultaneous drive failure is extremely unlikely in a conventional home server; with power shared with CPU and integrated controller, the only way you get to simultaneous failure is with simultaneous drive hardware failure which has a probability so low it's not worth considering (in a HOME environment!)....clipped...
But I think all this discussion, both the failure mode and my recovery approach, is beyond a typical home installation....
I would disagree strongly with the "not worth considering", as drives in the same case experience more or less the same heat, vibration, power on hours, power fluctuations, and other environmental factors. All these are contributing factors towards drive failure. If two or more drives are the same brand and model from the same production run, then that would also contribute towards them being likely to fail at close enough to the same time as to preclude a replacement having finished. Also note that, any kind of recovery puts additional strain on the remaining drives.
Yes, I've seen multiple failures over the course of two days at my own home. Even those of us with redundancy at home don't always instantly notice and replace drives that are experiencing issues immediately.
All RAID or other redundancy is beyond a typical home installation.
September 3, 2009 at 11:49 am
Nadrek (9/2/2009)
Pure software RAID-5/6 with Linux is very cheap, and more complex than I personally want to spend time on.Pure software RAID-1 with Linux I'm planning on playing with as a stopgap until I can budget a true RAID-6 replacement.
Intermediate firmware/software RAID I've never messed with at all.
Pure hardware RAID-1 isn't too expensive, and probably isn't too difficult.
Pure hardware RAID-5 I've been using at home since about 2002, and it's really quite easy to set up if you simply walk through the text based BIOS menus on the card. It also cost perhaps $400 for the controller in that year, plus the cost of a set of ATA drives.
Pure hardware RAID-6 is my replacement plan, and it should be of similar relative simplicity, and cost on the order of $800 for the controller, plus $100 for an LCD faceplate, plus the cost of a set of SATA drives. Advantages will include regular RAID consistency checks, which the prior card did not have available.
Speed is almost completely irrelevant - 1Gbps Ethernet is the bottleneck for sequential access, and for the rare random access requirement the budget constraints of magnetic SATA drives constrain me; a local RAM disk, Intel X25-E, or OCZ Vertex would fulfill any major random access requirement I had.
This is about where I stand. I'm curious, however, where WHS comes in here. As stated several times, this is for home use, not work. I believe that most folks on here could figure out how to script backups to a RAID server, some more easily than others.
Personally, learning how to use Web Based Enterprise Management to script that kind of stuff is usually a learning experience and I consider it a fun. That said, it's not worth much to simplify the task. However, I don't really see WHS costing much, considering the combined cost versus the hardware cost is almost negligible.
What's more, it looks like I might be able to get a service that's similar to RAID hardware setup for less cost. What I'd really like to know is, why didn't MS make some of this available on the 2008 server? (Or is this a hidden gem in there somewhere?)
Mind, it wouldn't be enough to make me abandon my linux server, where I can setup email servers, DHCP servers, time servers -- well, you get the idea. But that wouldn't preclude me from buying a WHS for backups.
September 3, 2009 at 11:57 am
Sir Slicendice (9/3/2009)
True simultaneous drive failure is extremely unlikely in a conventional home server;
I wouldn't count on that. As studies have shown, the idea that drive failures are statistically independent is seriously flawed. The greatest contributing factor for hard drive failure is time in use, and since most folks create these redundant systems with brand-new drives and place them into service at the same time, the likelyhood of a simultaneous failure becomes a real possibility.
That said, however, I do think a much greater concern for most housholds (in general, not constrained to this list), is doing backups at all. Most folks never think about it until after they've lost all tax records and vacation photos for the past five years.
Viewing 15 posts - 106 through 120 (of 135 total)
You must be logged in to reply to this topic. Login to reply