lionelhutz

How can it hold parity data for every individual file? If it did that then it'd be the equivalent of a mirrored system which it is not (well, it could be with enough multiple parity sets). I have not read anything about it storing say an MD5 or CRC type of checksum number either. Peter

OK, I posted "edit" twice now which does not mean adding new files. I edited the last post hopefully to make this clear enough. OK, I went to the Flexraid forum and the author posted an answer to NLS's post that confirmed my "conjecture" is exactly what would happen. After my xml file editing I would have badly hosed one of the other drives if it happened to fail before the parity had been updated. It's a big if, but still, it could have happened. It's not quite as rosy as Flexraid just not protecting my updated files. The Flexraid website says "However, if you edited or deleted data, only the un-synchronized portion will not be recoverable on the failed drive." which really isn't true. The editing you do on drive "A" will cause some level of corruption on the recovered drive "B". I also understand that Flexraid is for "static" data but who never changes or sorts or edits or deletes some of their data? Finally, I also understand that the problem only exists until parity is updated so therei is only a relatively small window of time where a failure could be very bad. As I posted before, I'm bringing it up because it could happen and anyone who's going to use Flexraid should understand it could happen. One response on the AVS forum claimed that drive "A" could be edited to your hearts desire and drive "B" could still be fully recovered at any time which just shows there is a basic lack of understanding of what Flexraid can and can't do. Peter

Well, as I have already stated, I posted this question over on the AVS forum where the AUTHOR of Flexraid was posting and did not get an answer. I already know it's for archived data but even archived data can be changed every now and then. I ask the question because I want everyone who's reading the thread to know the limitation and any possible "gotcha's" of the software. So having an actual answer would be much much more helpful. Huh, the example was me deciding to do an edit on a drive of archived movies. The metadata files are small xml files describing the movies, one for each movie. They get added as each movie gets added which means they are physically spread out over the drive media. I was changing the files to modify the movie genres that were in the files, not adding these files all at once. I don't see how I could do something like this without basically losing my parity protection until parity was rebuilt. If this breaks parity more than the Flexraid web site lets on then I feel people should know. It might not seem like a big deal but if I lost another drive while doing something like that the data on that other drive would end up badly corrupted. Another point is I'd have no idea what files were corrupted. Basically, I have to go through the data and figure out what is good and what is bad. Sucks to do on a 2T drive. Peter

Because I really can't be bothered and I'm not installing some extra software I don't want to use. I had posted this question on the AVS Forum where the developer was posting and never got a suitable answer there either. I don't take anyone's word that it "works" and will recover the data. If this is true then show me how. Both unRAID and Flexraid are claimed to do the same type of parity protection. Well, if unRAID only updated parity once a day and I changed block 18,352 on disk1 right after the parity calculation then block 18,352 on any other disk would be recovered wrong if I had a failure of any other disk. That type of random data corruption is not really acceptable to me. I just ran some scripts on my movie metadata files one disk1 of my server. These very small files get added as movies are so I'd bet they are spread out over the physical drive capacity. I can imagine how much data damage that'd do if the parity was not updated real-time and I had to rebuild one of the other drives after a failure. :'( Just giving anyone who considers this as a data protection solution some food for thought. Peter

Nope, not clear at all. Still not a technical explanation of how changing data without updating the parity doesn't break the protection for some files. Restoring the data on DRU2 would use both the parity and DRU3 but DRU3 was changed without updating parity. How does it get around the fact that DRU3 and parity are broken for some of the data? It's not exactly RAID4 since RAID4 has striped data but a fixed parity location. If unRAID was RAID4 then the data would be striped across all data drives and the parity would be on the parity drive. unRAID reads a "Stripe" from multiple drives and then builds parity data from that. If Flexraid is protecting files (it can protect files, folders or drives according to it's web site) then it must read a "Stripe" from multiple files and then build parity from that. So, Flexraid must put a bunch of data sources in "parallel" and then build parity as it reads across the "stripe" created by these data sources. I'm going to use a simple example. Say I have 4 files, each 5gig in size. I use Flexraid to create parity. So, flexraid puts the 4 files in parallel and reads the "stripe" data from them to create the parity. Now, I change file 1 and before I rebuild parity I lose file 4. How can file 4 be rebuilt??? I guess my point is that any form of parity protection relies on reading multiple blocks of data and then creating a parity block that matches. You can lose any one block in that "stripe" without losing any data. So, how can changing one file not affect the parity protection for another file? The responses saying that you can change any data you want and still recover any other unchanged data due to a failure just makes no sense to me. It's the equivalent of turning off the unRAID parity and changing the data on disk1 and then saying that disk2 can still be recovered. Peter

Here's something I could not figure out with Flexraid and even the developer on the AVS Forums never really answered. Say I protect complete drives like unRAID does - I know Flexraid can protect directories too but I'd have a "movie" drive so I want it all protected anyways. So, I'm using Flexraid to parity protect 3 data drives. From what I understand, it basically does a stripe parity calculation. I then add a few big files to one drive and remove a few more on one drive. Before I rebuild the parity another drive fails. Here's the big question - would I not get data corruption on the rebuild drive in the same physical locations as the new files and possibly the deleted files on the other drive??? There were posts that this isn't possible yet no explanation of how it could be avoided. The claim was that you could change the data all you wanted on one drive and then recover another drive if it failed and that makes no sense to me. I can understand that it would be possible to change the data on one drive and then recover that drive if it failed and it would go back to the old data. Peter

Could you move the data to the cache drive and then set the shares (drives, split level, fill level etc) correctly and run the mover? If you have a lot of small files spread out all over the place I could see the moves to the cache to be much faster when sorting the data. Run the mover once you are done sorting for the evening. You could also copy and allow the mover to put it in it's new spot and then go back and delete the old files. Peter

Here's another open box available right now too; http://www.newegg.com/Product/Product.aspx?Item=N82E16813135076R Peter

The Corsair is $39.99 after rebate and has 6 SATA connectors. Corsair is 80plus certified. Your PS + splitter add up to $43.98. I never looked up the Sparkle before but now I see it has dual 12V rails which is not a very good idea for unRAID. If you add more drives you could overload one rail. Peter

I've been using one of these for over a year now and absolutely no trouble. http://www.newegg.com/Product/Product.aspx?Item=N82E16813135231 - $45 There is a $50 one that has a DVI port but is otherwise the same if you decide that DVI might be something you'd want for the future. It works good, no HPA problems, has a fairly low power draw (<50W idle) and is prettty fast. I would try to this supply before that one http://www.newegg.com/Product/Product.aspx?Item=N82E16817139008 It'll easily suppport 8 drives with the other hardware you've picked. Peter

Good to hear it's working. Once you have the drives in and parity is built and checked it'd be easy to test the other PCIe ports. Yes, do a parity build and then do a parity check the first time. Once a new drive is installed do another parity check each time. Just move one of those cards with drives connected from the PCIex1 to the x16 slot. When unRAID boots you may have to go to the devices page and re-select the hard drives so make sure you have a list of drives before starting this test. Then, start the array and do a parity check. If the drives do not show up or you start to get parity errors then you know the slot is very likely no good so stop it and put it back. If it works then we all know the slot is likely fine for a 4 port SATA card and this board will likely become a highly recommended board based on it's cost vs number of PCIe slots. I think that Gigabyte has responded that the slot was graphics only with other motherboards but it still worked so the only way to know for sure is to test. Peter

That's really odd. A x1 lane should support a drive at full speed and you list the one card with a single drive as being slow. You should not begin to see a slowdown until you try to access more than 2 drives at once. Peter

I think it applies to every allocation method. If you copy multiple files at once the most free might change after one or 2 of the files get to the server. Peter

What's the difference between level 1 and leaving split levels blank or set to 0? I probably should have put 0 above there, even though I believe in this case 1 would work as well. Basically, use level 0 and create your own directory structure first. Then, when you drop files into this user share into an existing directory the allocation method is used to decide which drive, of the drives with the required directory, gets the data. Peter

Move the data you can to the array. Put the new 1.5T as parity once it arrives. Build parity and then do a parity check to confirm it is OK. Keep the working existing data drives as is until you have the parity confirmed. After parity is confirmed and you have moved the data from an existing drive to the array, install that drive in the array and add it. Did the media/DVD and media/TV shares appear as seperate shares on a windows machine - did they actually share? You are correct that those are just subdirectories of a share and I'm not sure what happens when you do that. You didn't post how you want the data to be stored. What I mean is - do you want certain data stored on certain drives? If you want it to just fill the array however then use level 2. This means that a TV sub-directory and DVD sub-directory will go to every drive but any subdirectories below those will not split. Each TV series will stay on a certain drive and each DVD movie will stay on a certain drive (multiple file movies will not be spread out over a couple of drives). The only trouble you might run into is that a TV series fills a drive since the whole series is confined to a single drive once the root "series name" directory is created. You could also use level 1 and manually create DVD and TV sub-directories where you want those to go. For example, start with Media/DVD on one 1.5T and Media/TV on the other 1.5T. Then, create a new DVD sub-directory on the third 1.5T once it is installed. This way, you pick where the DVD and TV folders go. The Popcorn Hour requires a single share point, correct? Peter

You've got 7 things to connect wired. The router has 4 ports and that switch has 5 ports. You use 1 port each to connect them leaving 7 ports. So, that switch would work but it leaves you no extra ports for the future. Put some of the 100Mbps devices on the router and the Gige devices on the switch. I guess you could also leave your 100Mbps switch plugged in to expand the router a few more ports. Personally, I try to run with as few plugged in devices as possible just because it saves complexity and hydro. Peter

Once you set that +0.3V then you should have been fine. You can run with slower timings (5 instead of 4) and it will just be slower. So, it looks like the RAM is bad. Peter

Well I don't really care. I really don't have much further tolerance or respect for further complaint posts in the beta thread that was intended to address and help with technical issues. The people complaining about Tom not responding in quite a while and not yet fixing the problems in the beta release have repeatedly been asked to start a new thread to discuss that issue so that the beta thread isn't swamped in non-technical posts. Since Tom not using the forums is more of a forum issue the most suitable spot for such a thread is here. In case anyone doesn't know, Tom decided that the next release would be something along the lines of having an open interface using a built in full web server. This is a major re-write of some parts of unRAID. I wouldn't even be surprized if he's working with some of the people here but they're silent about it too because mentioning what is coming would just start more speculation and conjecture and complaining about not yet having something. So, since unRAID has issues how about spelling out what is wrong with 4.4.2 since that is the last "finished" release? The 4.5 beta6 is a beta which is expected to have problems so when I find problems with it I don't call unRAID broken or unsuitable because it is expected. Also, if you were running a business would you appreciate a bunch of people going on about how you're such a bad business operator? I bet you'd especially like that if they do such without bothering to contact you with their issues. Unless I missed it, I did not see anyone post where they had called or emailed Tom. I have tried to not make any comments on the how's or why's or when's of Limetech operations. How Tom runs his business is his own business. If I decide I don't like Limetech and unRAID as a company/solution then I am free to go somewhere else. Similar to the Seagate drive issues, I just post about the issues they have had without feeling the need to go on about the company. All companies do things you don't agree with. I bought unRAID to use for network connected storage. It works perfectly for that purpose. I remember that it did crash on me once with no real reason in the last year or so but that happens with any computer. I'd like to be able to easily run other applications on unRAID instead of using a second Windows machine (which is already easy to run equivalent applications so I don' t want to spend much time figuring it out on unRAID) so future development would be nice. However, it does the network connected storage thing perfectly fine and has met if not exceeded my expectations there so I just don't have any reason to complain. Now, if others have issues with 4.4.2 or previous production versions they should be contacting Tom. If you feel a broken feature of 4.5beta6 is keeping you from purchasing a new license then you should also contact Tom and discuss the situation to him. Peter

OK, so how about you go here and leave this thread alone. http://lime-technology.com/forum/index.php?topic=4397.0 This thead needs to be split so there is the new no Limetech response thread for everyone to move to. Peter

Did youi contact Tom DIRECTLY? unRAID 4.4.2 is the last production or final release. Are you having problems with it? I just switched to this beta but was using 4.3.3 without any issues. Complaining about problems with a beta release like you are owed fixes is where I have a huge issue. It is a test release. Go to the Limetech main web site. Exactly what part about the unRAID descriptions there has not been met by 4.4.2? Exactly where does it say that Limetech promises to release new features and fix bugs in said features in a timely manner. I won't deny there are issues with this release. Tere are some issues with the production releases but more along the lines of the interface not accounting for certain conditions then actual bugs or broken features. I'm just hoping Tom is working on a new version 5 of unRAID that opens up the UI. Then, some of the great support people here (like Weebo, Joe, BubbaQ) will likely blow you away with their add-ins you can get. Personally, I'm hoping a web based torrent server will be easy to do among others. Everyone else, go ahead and make yourself feel better by complaining here. Peter

Very good work as usual Rob, that sums it up very well. It really reminds me of the Satellite hacking boards of earlier years where the users would bitch and moan about no new fix so badly that the developers would just finally declare "Screw You" and leave. It's funny how so many people think they are entitled to get things for FREE. Getting something extra for free is a privilege, not an entitlement. To me, any free software update after the point when I purchased it is a big bonus, not a requirement. "I purchased a product that does exactly what it was advertised as doing when I purchased it but the developer still owes me more!". Seriously? You won't get any major free updates from Microsoft, or many other software companies for that matter. Having said that, I look forward to the possibilities that version 5 will bring and I do hope that it is produced. Peter

This motherboard would give you much faster disk access speeds because of the dual PCI-X busses so add 2 of the AOC SAT2-MV8 (one in each buss) and you are at 20 ports total. http://lime-technology.com/forum/index.php?topic=2697.0 Or find a motherboard that has 4 PCIe x4 slots to hold 4 of these and also has 6 onboard. http://lime-technology.com/forum/index.php?topic=2820.0 Peter

As I posted, I don't know what you are trying to store. I have a number of shares. Things like TV and Movies follow fairly defined folder structures so I can use a split level of 1 and be quite safe. In fact, I would not want a specific movie directory or TV series directory to split. If the specific movie directory split then that would require that multiple drives were read to access that movie. Kind of the same thing for a TV series. Other random data might not behave as well and require a higher split level. And yes, there will be a disk full error if the settings lock the directory to a specific drive that is full. I personally don't like or see the need for one big share. I have a few like Movies, TV Shows, Pictures, Applications, etc. There are different requirements for each so I do that to keep them sorted. But, what you post is OK if you just want one big share and don't want to keep any specific organization. I believe that a split level of 0 will also allow the splits to happen wherever. Peter

Good job going for it and getting the info. Hope they work out and you can build some cheap storage. Peter

Depending on the data you are storing you may want to use a split level of 1 or 2. This would make sure that data in subdirectories is all kept on the same drive. I don't know enough about the data to recommend something though. Peter