NASLite Network Attached Storage

Before v2 raid was a no go, but a lot of users wanted it and its was added to the v2 version. those that asked for it more then likely knew the costs of the cards or even had a card laying around.

there are other os that do the software raid bit, its not the developers fault that the cards cost as much as they do, but they gave people a choice shoudl they take it. naslite was designed to be fast and dependable and i guess software raid would have impacted on that.

Let me say it again ...



As I said before, RAID has an entirely different purpose to backup, and perhaps I should also say, having a backup does not eliminate the need for RAID - your solution with duplicated NAS servers is a backup solution, however if you were using it in a business environment and a drive in the primary NAS server failed, your business would suffer down time, you would also have to deal with possible data loss, depending on how frequently you were backing up to the secondary NAS server.

Just as sjdigital would have to buy two "buff boxes" to have his data and a backup, if he were using NASLite with or without RAID, he would need to duplicate it.

The original discussion was based on the additional cost of RAID hardware placing NASLite into competition, pricewise, with full featured, commercial NAS solutions - and if we remove eBay as an option, the math presented by sjdigital is unargueable.

If Server Elements were to round out the NASLite feature set, adding UPS support, user security - especially the ability to integrate into a Windows domain and a decent RAID management toolset, NASLite could be moved from a niche market enthusiast's product to a SOHO/SMB contender.

I have no doubt that there are challenges to making my suggestion a reality, but surely the work involved in meeting these challenges will be rewarded by increased sales in a significantly larger mearket segment.

I understand the difference between Raid and Back up, But the orignal poster was questioning the costs of adding a Hardware raid card to a Nas box over buying a Buffao box. and that the Buffao box was better value for money. The solution would be that Nalite went software then you could compare both. But you can't really compare one with the other when in one you have a expensive Hardware card in one on none in the other.

The math presented is flawed to really compare you have to do like for like.

Again its not server elements fault the price of cards is high.

Using Raid at a business can still encounter a failed disk and screwed up array, just as much as my back up solution, but I could purchase a new hd that day and have one nas back up and running in a short time.

It's interesting to see that my original query (which as I'm sure most of you will have realised by now was predicated on almost total ignorance of the subject!) has prompted quite a debate about the role of RAID in the context of the NASLite concept.

But as I'm really only looking at it from the home user point of view, perhaps on the basis of ex abundante cautela the answer is to set up a 'consumer' NAS box (not necessarily RAID) with media server, shares, access control etc., which in turn is set up to back up to an industrial strength NASLite RAID 1 box. Not the cheapest solution perhaps, but it might be pretty bullet proof.

What say ye?

At the end of the day its down to the user, if they want hardware raid then they have to part with the cash, if they want a cheaper solution then they also have that option. if naslite doesnt do everything a user wants then maybe they have the wrong OS? also I guess its how important your data is to you and how much your willing to pay to protect it.

Actually, if I might disagree, it's not (or shouldn't be) a matter of wanting hardware RAID in preference to software RAID. Surely it should be a matter of what degree of redundancy and reliability is required and then selecting the most cost effective solution to achieve that. Form follows function, and all that!

So if I simply want to back up my notebook as I travel, then a regular NASLite box, accessible via VPN from wherever I am would suffice. But if I need more than just a back up location, e.g. I want to run a medai server, then it will also have to be backed up. In which case it may make more sense to have dedicated media server box which is itself backed up to the NASLite box

The decision to purchase a hardware raid card is no different than buying a gigE network card or a bigger monitor.

If I need faster file transfers, I'll get a gigE card, If I think I need to see the NASLite-2 status pages without having to scroll down, I'll get a bigger monitor. If you want large amount of storage, that took you countless man hours to aquire and can't afford to lose it or have down time, the price of the raid card becomes meaningless, it's a form of insurance. The point here, if you don't need or want it, don't buy it then.



I don't see the point there either, no one's married to NASLite, you could take your hardware raid formatted under ext2/3 and put it in a Windows box, a OS X box, or another unix machine. If someone outgrows NASLite, hence the name there, they have the option to move on with thier data.

ok i'll put it another way,

At the end of the day its down to the user, if they want software raid then they have that option, if they want a more expensive solution then they could go the Hardware Raid route.

But it's not a matter of wanting "software raid" or wanting a "more expensive solution". Surely it's a matter of what degree of redundancy/security are required and then selecting the most cost effective option to achieve that?

hence At the end of the day its down to the user.

Exactement!

Here are some basic guidelines that I prefer to observe:

I'll consider RAID when I can't afford any downtime during operation. Cases where people have multiple open files that represent hours of work can arguably justify the use of RAID. In other words, the risk in a mostly dynamic filesystem can be reduced by the use of RAID.

On the other hand, when hosting a filesystem that is used primarily for storage of content that does not change very often, periodic mirroring of the content can be sufficient to ensure redundancy. RAID in such scenarios is not necessary since the mirroring does not need to occur in real time.

There are other reasons one may elect to use RAID such as performance, but in the context of NASLite that should rarely be necessary.

I know that this is often very subjective, but if you base your approach primarily on your task requirements, it is often very easy to determine if RAID is necessary or not. In the context of NASLite, all you have to ask is how much data will be lost if I have to revert to a 24 hour old mirror. If the loss does not justify the cost of RAID, then the daily mirroring option will be sufficient.

I do understand that the above is rather oversimplified, but I hope that it makes sense and gets the point across.

Excellent advice methinks, and a sensible conclusion to this thread. Thank you.

Although a bit late in commenting, here's my 2 cents.

I think the discussion was based on the "high cost" of a RAID controller.
The 200 pounds price is not realistic. It is based on street retail prices which are prohibitive for most home users.

I had the same thoughts and during the summer I managed to get not 1 but 2 raid cards via ebay. One 4 port PATA adaptec 2400 for 55pounds shipped from the UK to Greece and one 8 port PATA 3ware 7500 for $150 shipped to Greece. I think if one is persistent and patient, they can find a good deal on ebay.

Based on the above prices I can say that a hardware raid card and NASlite are very cost effective.
Here's the cost of my NASlite V2 server with 2.5 TB of storage.

PC Case, PSU, Fans, Disk Bays, MoBo, CPU, RAM = All new $250
3ware 8 Port RAID controller = $150
8 x 300Gb PATA = 8 x $120
2 x 300Gb SATA = Already owned.
2 x 80Gb PATA = Already owned.

Total $1360 of which $960 is the cost of disks.

This could be made much cheaper if I used an existing PentiumII PC and did not opt for disk bays with fans which added about $100 to the price.

Overall, I cannot understand how it can be more effective to have dual servers since obviously most of the cost comes from the disk drives themselves.

I wouldnt say its effective over raid, and my coments were regarding the original post where by the Buffao guys recommend a second server to back up the first, so in effect you have to purchase two buffao boxes.

My suggestion was to have two Nalite boxes without raid and to mirror the drives to another nas box, you could if you wish have all the drives on the local machine and do away with the second box. the cost of this is a lot less then buying two buffao boxes. as you point out you can purchase radi cards on ebay, but should a card fail or be failing? wheres you data gonna be when it dies. looks like you would have to locate a second card and more expense.