RockDawg

Just to be clear when you say "7. Run parity check on new hardware", you mean rebuild the parity, right?

I have one more question. Since I will be combining two separate unRAID servers into one, is the best course of action to run parity checks on both servers before dismantling and if both return no errors then put all the data drives into the new build and just rebuild parity? Are there any other things I should do or be mindful of?

I am looking at the Supermicro X9SCM-F-O at SuperBiiz, but the only ones they have in stock are bulk packed instead of the retail pack? I've purchased many bulk CPU's in the past, but never a motherboard. I don't care about the cables and what not, but I hate not having an I/O shield. Do the bulk MB's come with the I/O shield?

brian89gp - Thanks for the feedback!

Thanks for the info. In my casual perusing, I didn't see any make or model of controller(s) listed on NewEgg or Supermicro's site and I assumed multiple controllers since there are different speeds for different ports.

I am seriously thinking about building an Atlas clone and virtualizing with ESXi. Right now I have two unRAID servers on 24/7 (one with 10 data + 1 parity and the other with 9+1) and my desktop PC running 24/7 to handle SickBeard, CouchPotato and Sab. The main reason for the desktop staying on is to update the XBMC library after SickBead completes a download. So I am looking at 3 computers on constantly that I could whittle down to one by virtualizing. I’m not Linux savvy and the most I’ve done with virtualizing is install and play with Ubuntu and Mint in VirtualBox a couple times. My initial intentions would be to run unraid, and Windows 7 for SB,CP and Sab. I’ve been researching this for a couple weeks now and I have a few questions (forgive me if some of these have been previously discussed. It’s been a lot to absorb and I’m sure I’ve forgotten some things I’ve already read). 1. Just so I totally understand this, I assume that the datastore for the Win7 VM would essentially be my C: drive and contain all the data for that VM? 2. I’ve read about using pass through for the HBA controllers (I have two IBM M1015’s) and I understand that you have to pass the entire controller to the OS and that’s the only OS that can then access it. It looks like the X9SCM-F-O has two separate onboard SATA controllers. Can I pass the 4 x 3Gbps controller through to unRAID and use the of the 2 x 6Gbps controller for the datastores ? 3. I’ve read some people talk about using a separate NIC for each VM, but it seems more people are doing the virtual switch instead. (I assume that means that ESXi is handling the network traffic management via a single NIC?). Which is the best route and why? 4. A big concern for me is the hoops you have to jump through to get a UPS working. I currently have an APC UPS and it handles both servers without issue. I’ve read about people using scripts to shut down the different VM’s and then being concerned about whether or not the UPS will then shut itself down. I haven’t seen any follow ups on that. Does the UPS shut itself down and everything power back up on it’s own once power is restored? I don’t really know how to write a script, but are the ones necessary very complex? I really don’t like the idea of being without the UPS. That’s all I have for now, but I’m sure I’ll end up with more as I get further into this.

I just had this happen to me a couple minutes ago. I was listening to some MP3's that are on my unRAID server and I started to rip a DVD to that same server and as soon as the rip started the MP3 playback paused for about 6 seconds and then continued. I checked the log and there isn't a single entry anywhere near the time that occurred. It's not a network congestion problem either because I had been listening to MP3's for the previous hour or so and I during that time I copied a movie to a different unRAID server on my network and MP3 playback never missed a beat. I would love to see this get resolved as it's annoying. I am running 5.0 beta 12a EDIT: I have ripped another 4 DVD's and it happens every time..

I've noticed glitches like this over different versions of unRAID for quite some time. Some examples are: 1. MP3 stream skipping while simply navigating to an unRaid drive in Win Explorer 2. Video stream completely stopping during a file transfer to a share on the server When I get home I try a couple of these and catch the log.

I haven't been getting any graphics either since the migration. Is this still not fixed?

Depends on what you mean by "continues to be absent". I've been using unRAID since 2005. Shortly after he first announced it on AVS Forums and before he had this site. He has a history of doing this going back to January of '06. Here's a post in that AVS Forums' thread where people start to get concerned: http://www.avsforum.com/avs-vb/showthread.php?p=7122302#post7122302 This has happened numerous times throughout the years. In fact if history is any indicator, it's a veritable certainty that he will disappear again. The problem here, as usual in this world, is that both sides have a point but the extremists in both camps are way off the mark. You have the PO'd few who feel they are owed something more and come out and post abrasive comments. Then you have the other side of a few fanboy loyalists who will forever deny that there's any problem and say the others should just be quiet and be happy with what they have. IMO both are wrong. Tom ultimately owes us nothing more. He has a stable product that performs at least as well as advertised and is offered at a fair price. If he were to never release another update or upgrade, we would not have a valid complaint of being ripped off or undersold. With that said, consumers are not wrong for having their expectations. In the computer world things move quickly and people are used to that. In this digital age with email, forums, texting, Twitter and Facebook it's hard for people to buy into the fact that someone is so busy that they can't be heard from. Who doesn't have at least some down time to make a quick post at some point during most days or at least the week? It's like when you see an old friend you haven't talked to in a while and they say, "Sorry I haven't called, but I've been sooo busy." It's BS and we all know it. And that's just about that concern with the lack of info from Tom. That's not even getting into the years it's taken to get to 5.0 (and we're still not there yet). I think we all have a certain love or appreciation for unRAID. That's why we're all here. Personally, I fall into a middle group. I love unRAID and think it's an excellent product as-is. BUT, I would like to see development move a little quicker. Any development. If all this wait is for a stabilized kernel with support for newer hardware, then why not just say that and implement an updated GUI for 4.x and above while we're waiting. Something like speeding_ant's Simple Features GUI would be great (great work by the way speeding_ant!). Give us some built-in support for email notifications and UPS. I think something like that would have gone a long way in tiding people over until 5.0 does come out. Whatever happens, I'm here for the long haul because ultimately, unRAID just works. Just my $0.02.

Thanks for pointing this out. I assumed that if I excluded a disk from a share then it's existing data would not show in the share either. I didn't realize that setting was for writes only. Seems like that should do the trick.

I just thought of another reason why the min free space setting is probably not the best option for what I'm wanting to do. You can only set a single value, but if you have varying sizes of disks, this would cause a problem. For example, my smallest drive is 500GB and my largest is 2TB. To stay under 99%, but still fill the disk as much as possible (say 98% full), I would need to set free min space to ~10GB for the 500GB drive, but the 2TB drive would need to be set at ~41GB. If I set the min free space setting in the user share to 10GB, I over fill the 2TB drive and performance suffers. If I go with the 41GB setting, every drive stays under the magical 99% mark, but I end up with a fair amount of wasted space. For my total array I estimate I would waste ~175GB if I went with the 41GB setting. Again, I think this would be better as drive setting rather than a share setting. Feel free to point out if my thinking is flawed.

I didn't realize that was per disk. Stupid I guess, but I thought that was overall. Still, it would still be much easier to simply mark a disk and it be be excluded from all share writes rather than have to go through each share and set the minimum free space. Not to mention that that could sometimes change, too. If a disk is full, it's full. This should be a disk level option, not share level.

I only recently stumbled upon the "Avoid filling drives >99%" entry in the Wiki, but has impacted me greatly. I used to get a ~45 second delay sometimes while browsing my unRAID servers via my Win 7 desktop. After redistributing some data to keep all my drives <99%, I can now navigate them much quicker. With this in mind I was thinking how it would be nice if we could mark a drive as full and no longer write to that drive via user shares so as to avoid filling it above 99% and incurring the potential slowdown. I would think you could still allow writes to the disk so as to allow an easy manual override. I'd like to just mark a drive and know that Sick Beard will no longer write anything to that disk while writing to my "TV" share. On a side note; why is the only Feature Request forum for 4.x and closed? I would think that a FR forum should always be open and available. Not a biggie. I never thought about it or looked for it before, but I found it a bit odd there was no open FR forum/thread.

Thanks a ton! I tried searching for various terms including the word "bracket" on ebay, but couldn't find anything. I appreciate your help.

I checked on ebay and they only had a low profile bracket and I don't know of any places around me that might have anything like that. That's why I posed here.

I need a normal high profile bracket for my M1015. I could've sworn I had a spare card with on that fit, but I can't seem to find it now. Does anyone have one they are willing to sell or know where I might be able to find one? The holes are spaced 2.125 inces (54mm) apart.

After the preclear completed without issue, I assigned the drive to my array and it has been fine since. No more of those errors in my log. It would still be nice to know what these errors meant for future reference.

Yah sorry. The log was too big to attach and too big for pastebin. I had to find an alternative. Here it is: http://pastesite.com/31290 That drive is new and I am currently preclearing it.

I am running 5b12A and I am getting a bunch of these repeating entries in my log: Feb 3 15:30:53 Tower2 kernel: sd 0:0:0:0: [sdj] CDB: cdb[0]=0x85: 85 06 20 00 00 00 00 00 00 00 00 00 00 40 e5 00 Feb 3 15:30:53 Tower2 kernel: scsi target0:0:0: handle(0x0009), sas_address(0x4433221101000000), phy(1) Feb 3 15:30:53 Tower2 kernel: scsi target0:0:0: enclosure_logical_id(0x500605b001f27e80), slot(2) Feb 3 15:30:53 Tower2 kernel: sd 0:0:0:0: task abort: SUCCESS scmd(c30699c0) Feb 3 15:31:02 Tower2 kernel: sd 0:0:0:0: attempting task abort! scmd(e8d1dd80) Feb 3 15:31:02 Tower2 kernel: sd 0:0:0:0: [sdj] CDB: cdb[0]=0x85: 85 06 20 00 00 00 00 00 00 00 00 00 00 40 98 00 Feb 3 15:31:02 Tower2 kernel: scsi target0:0:0: handle(0x0009), sas_address(0x4433221101000000), phy(1) Feb 3 15:31:02 Tower2 kernel: scsi target0:0:0: enclosure_logical_id(0x500605b001f27e80), slot(2) Feb 3 15:31:02 Tower2 kernel: sd 0:0:0:0: task abort: SUCCESS scmd(e8d1dd80) Feb 3 15:31:26 Tower2 kernel: sd 0:0:0:0: attempting task abort! scmd(c5168f00) Feb 3 15:31:26 Tower2 kernel: sd 0:0:0:0: [sdj] CDB: cdb[0]=0x12: 12 00 00 00 24 00 Feb 3 15:31:26 Tower2 kernel: scsi target0:0:0: handle(0x0009), sas_address(0x4433221101000000), phy(1) Feb 3 15:31:26 Tower2 kernel: scsi target0:0:0: enclosure_logical_id(0x500605b001f27e80), slot(2) Feb 3 15:31:27 Tower2 kernel: sd 0:0:0:0: task abort: SUCCESS scmd(c5168f00) Feb 3 15:31:49 Tower2 kernel: sd 0:0:0:0: attempting task abort! scmd(c9f746c0) Feb 3 15:31:49 Tower2 kernel: sd 0:0:0:0: [sdj] CDB: cdb[0]=0x12: 12 00 00 00 40 00 Feb 3 15:31:49 Tower2 kernel: scsi target0:0:0: handle(0x0009), sas_address(0x4433221101000000), phy(1) Feb 3 15:31:49 Tower2 kernel: scsi target0:0:0: enclosure_logical_id(0x500605b001f27e80), slot(2) Feb 3 15:31:50 Tower2 kernel: sd 0:0:0:0: task abort: SUCCESS scmd(c9f746c0) Feb 3 15:32:24 Tower2 kernel: sd 0:0:0:0: attempting task abort! scmd(f3974840) Feb 3 15:32:24 Tower2 kernel: sd 0:0:0:0: [sdj] CDB: cdb[0]=0x12: 12 00 00 00 24 00 Feb 3 15:32:24 Tower2 kernel: scsi target0:0:0: handle(0x0009), sas_address(0x4433221101000000), phy(1) Feb 3 15:32:24 Tower2 kernel: scsi target0:0:0: enclosure_logical_id(0x500605b001f27e80), slot(2) Feb 3 15:32:24 Tower2 kernel: sd 0:0:0:0: task abort: SUCCESS scmd(f3974840) Feb 3 15:32:47 Tower2 kernel: sd 0:0:0:0: attempting task abort! scmd(f31c6f00) Feb 3 15:32:47 Tower2 kernel: sd 0:0:0:0: [sdj] CDB: cdb[0]=0x12: 12 00 00 00 40 00 Feb 3 15:32:47 Tower2 kernel: scsi target0:0:0: handle(0x0009), sas_address(0x4433221101000000), phy(1) Feb 3 15:32:47 Tower2 kernel: scsi target0:0:0: enclosure_logical_id(0x500605b001f27e80), slot(2) Feb 3 15:32:48 Tower2 kernel: sd 0:0:0:0: task abort: SUCCESS scmd(f31c6f00) Feb 3 15:33:59 Tower2 kernel: sdj: sdj1 What do they mean? Full syslog is attached.

Thanks. I overlooked the bus speed difference. What do you mean by suports larger hard drives? According to the first page, both now support 3TB with the right unRaid version. Are you saying the BR10i won't support anything larger?

For the purposes of unRaid is there any functional differences between the M1015 and the BR10i? I'm guessing that once we flash these to just be HBAs, they all pretty much perform the same. Is that correct? The M1015 seems to be going for $70-80 on ebay and the BR10i can be had for as little as $40. Not being real familiar with this kind of tech I'm wondering if there any reason to spend the extra money on the M1015.

I think I finally got everything sorted out on both systems. I bought two new hard drives. They are precleared and installed and parity and data has been rebuilt all seems normal now and I'm getting write speeds ~30 MB/s on both drives. After I pulled both questionable drives, I ran them both through WD's Data Lifeguard diagnostic software. They both returned the same results. They passed the short test and then completed the long test with some bad sectors that needed repaired. When I instructed it to repair the sectors, both drives failed with a nondescript error message that said there was an error and the repair could not be completed (or something like that). I RMA'd both drives. I'm still wondering if the power supplies indeed caused this. Either way, I'm glad I now have good power as it gives me peace of mind. Anyhow thanks for all the help dgaschk and Johnm!

I ran down and bought another Corsair GS600 PSU and replaced the one in the server. It turns out it is an Antec EarthWattsd 430W. Dual 12V rails at 14A a piece. My ignorance strikes again! Whether or not the PSU was the sole problem in either system, I still feel better knowing I now have proper power to both systems. I had a spare 2TB drive that had preclearing and it's now finished so I also decided to just pull disk7 and replace it with this new one. Once everything seems normal on this system I'll run preclear a few times on the drive I took out and see what that tells me. I'm still miffed at the first system. It's chugging along a parity rebuild at 4 MB/s now. This could end up taking 3-4 days at this pace. Is it possible the lack of proper power did irrevocable damage to the drive?

It's been rebuilding parity for ~21 hours now and it's only 24% done it's been averaging 7 MB/s everytime I look at it. Shouldn't the speed have increased if the PSU was the sole problem? Here's a link to the log (too big to attach). This server is Tower2. http://pastesite.com/30739 What's really strange is I have a second unRAID server (Tower1) and it's started acting similarly. It has a drive that is getting really slow write speeds (5-6 MB/s) and the SMART report again shows some pending sectors ( and zero reallocated. How common is it to see pending sectors, but non reallocated? This makes three drives across two servers. Once again, when I try to write something to this drive from my Win 7 desktop I get the same error as I did on the other server, "There is a problem accessing \\Tower1\disk7. Make sure you are connected to the network and try again". The SMART report on the parity drive of this system seems OK. Only disk7 seems to be the problem. It's showing the following lines in the syslog sometimes: Jan 21 15:51:03 Tower1 kernel: ata9.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen Jan 21 15:51:03 Tower1 kernel: ata9.00: failed command: SMART Jan 21 15:51:03 Tower1 kernel: ata9.00: cmd b0/d0:01:00:4f:c2/00:00:00:00:00/00 tag 0 pio 512 in Jan 21 15:51:03 Tower1 kernel: res 40/00:ff:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout) Jan 21 15:51:03 Tower1 kernel: ata9.00: status: { DRDY } Jan 21 15:51:03 Tower1 kernel: ata9: hard resetting link Jan 21 15:51:09 Tower1 kernel: ata9: link is slow to respond, please be patient (ready=0) Jan 21 15:51:13 Tower1 kernel: ata9: SRST failed (errno=-16) Now I haven't looked and I don't remember what PSU is in this system, but I'm sure it's not up to snuff either as far as being single rail and having enough amps. But it seems so strange that all this would happen to both servers at the same time. When they've both been running fine for so long. I appreciate all the help and I'm sorry to require so much of it. I'm just lost on trying to solve this.