March 2, 201016 yr Short: I've been flip-flopping on WHS vs unRaid and decided to to unRaid first, maybe add a WHS later if "needed". Keeping it cheap; using as much old hardware as possible until I can afford a Norco case. I have no real experience with Linux of any kind. Played with a few distros for a short time. My main concern appears to be my current and potentially future motherboard options aren't on the hardware compatibility list for unRaid. Current board: ASRock K7S41GX Future board (currently in my htpc: ECS KA3 MVP). Neither of those boards are on the supported list for the Promise SATA300 TX4 I just picked up either. My concerns here are that a) in my current board it will work fine with unRaid and b) it will play nice with the onboard controllers (ATI sb & jmicron raid) on the ECS board. How can I test for this? If I open it, I can't return it lol. Even so, I can't just see if it works in a windows environment and expect that it will work with unRaid. I guess if I do, it will at least confirm that the hardware itself will/won't play nice together. My first question is, I ended up paying $85.00 (Canadian, tax in, local shop) for the Promise TX4 which was kind of hard to swallow. I've searched the forums and it appears to be a popular/well liked card but at that price, I could pick up several syba cards! (In fact, I originally picked up a 2 port card for $12, but 2 ports aren't enough). I've seen enough people using Syba/other SiI3114 based cards that I'm thinking I should go that route instead. The one thing I haven't seen answered is: do the SiI based cards play nicely with other cards, in general? If so it only makes economical sense to ditch the Promise card unless there is good reason to keep it. Edit: also considering: http://www.vantecusa.com/front/product/view_detail/391 Anyway, this week I plan to start playing with unRaid on the ASrock board. I will update this thread as my testing continues. Thanks!
March 2, 201016 yr This board: ASRock K7S41GX will be a toss up to if it works or not. It looks like it uses SiS chipsets for almost everything which are listed NOWHERE on the Hardware Compatibility Page. You could always use the free version of unRAID to boot the motherboard and see if the NIC gets an IP address and how well it works. The webGUI should be reachable if the NIC is working and getting a connection. This board: ECS KA3 MVP would be a better contender for now. The extra PCIe slots will come in handy later!! The first NIC port looks to be an odd controller but the second uses a realtek chipset and should work if you disable the first in the BIOS. The x16 and x1 PCIe slots will allow you to expand much more later on and not feel to much slowdown when doing a parity check. This card has not been tested here on the forum but it will fit into the x1 PCIe slot and be much faster then trying to use a card that is only PCI based. It uses a Sil chipset and the card can be flashed with the non RAID BIOS, which should in theory, allow it to work in an unRAID system. Neither of those boards are on the supported list for the Promise SATA300 TX4 I just picked up either. My concerns here are that a) in my current board it will work fine with unRaid The only way to know for sure is to prepare a flash drive and boot it on the board. If it boots and the NIC gets an IP address you may be good to try it. If you do get it to boot I would suggest posting the syslog so some of the other members can take a look through it to see if we can find anything wrong. You can find direction on how to get a syslog in the troubleshooting section of the unRAID wiki. b) it will play nice with the onboard controllers (ATI sb & jmicron raid) on the ECS board. Any sort of RAID is pointless in unRAID so do whatever you need to to disable it. Set the controller to AHCI (if possible) and let it do what it likes. My first question is, I ended up paying $85.00 (Canadian, tax in, local shop) for the Promise TX4 which was kind of hard to swallow. I've searched the forums and it appears to be a popular/well liked card but at that price, I could pick up several syba cards! (In fact, I originally picked up a 2 port card for $12, but 2 ports aren't enough). I've seen enough people using Syba/other SiI3114 based cards that I'm thinking I should go that route instead. The one thing I haven't seen answered is: do the SiI based cards play nicely with other cards, in general? If so it only makes economical sense to ditch the Promise card unless there is good reason to keep it. That is really up to you. If you went with the second board and got the SATA controller card I linked to you might get it working. There might be a little messing with it at first, but my gut feeling that combo should work. Edit: also considering: http://www.vantecusa.com/front/product/view_detail/391 Any card with that many ports on the PCI bus is going to be painfully slow when doing parity checks. If you can avoid the PCI slots (NOT PCIe) for SATA controller cards then I would. Well, hope that helps a little bit, and good luck on your build.
March 3, 201016 yr Author Wow, thanks for the detailed reply prostuff! Here we go: trying unRaid out on the ASRock mobo with nothing in the system just yet. Got as far as loading bzroot.... taking a while to load and welcome to unRAID! And... web GUI is accessible. Ok cool, now time to test some drives in it. I'm going to start with an elcheapo Syba PCI card based on the SiI 3512 chipset and see what I can see. I have some other comments to yours but for now I just want to keep playing I'll try and post a syslog a little later. EDIT: I've attached the syslog from when I had both the syba 3512 pci controller and the promise sata300 TX4 as well (which I think I'm keeping). Let me know if anything looks wonky in there. Pretty much jibberish to me at this point. Ok, now to address some of your points that you amazingly took time to make (thanks again). So from what I can tell, it works no problem with the ASRock board. I'm happy about this because the ECS, while eventually slated to replace the ASRock, will not happen until later this year when I can afford to upgrade my HTPC's components. For now, my goal is to get a *working* unRAID setup. I know it's not going to be the fastest and I know I should ultimately be looking at PCIe controller cards but this box isn't going to have that right away. For speed, I'm keeping my Raid 5 array in my HTPC for now and intend to use the unRAID box for less performance oriented things until I get a better feel for it. Tomorrow night I'd like to put the ECS board through its paces and see if I can get that "odd" gigabit NIC to work. Regarding the Vantec card, it has 6 ports and is marketed as such but really only uses the SiI 3114 chip which is a 4 channel card. There's a jumper on the card to activate all 4 internal or 2 internal, 2 external. I liked that feature as again, for starters, I may need to use my external eSata enclosure or two outside of my puny case until I can afford a Norco 4020 or 4220 case. While longer term testing would be required to be conclusive, it appears that the Promise internal ports have no problem running my external enclosure even though I've seen reports on newegg of this being flaky with this card. I think I'm keeping it. While the SATA II spec may not really make any noticeable difference over say a SiI 3114 based card, I like the fact that it supports NCQ and the like. So now that I have it up and running, what do you suggest I try and to to ensure it will be a robust, platform? syslog-2010-03-02-Slipstream.txt
March 3, 201016 yr Wow, thanks for the detailed reply prostuff! Your welcome, feel free to ask more and we will try to help! Here we go: trying unRaid out on the ASRock mobo with nothing in the system just yet. Got as far as loading bzroot.... taking a while to load and welcome to unRAID! And... web GUI is accessible. Your report of the NIC being recognized looks good. unRAID loaded the sis900 driver Mar 2 22:02:09 Slipstream ifplugd(eth0)[1021]: Using interface eth0/00:0B:6A:DC:A6:CA with driver <sis900> (version: v1.08.10 Apr. 2 2006) and your board continued setup from there. For now, my goal is to get a *working* unRAID setup. I know it's not going to be the fastest and I know I should ultimately be looking at PCIe controller cards but this box isn't going to have that right away. For speed, I'm keeping my Raid 5 array in my HTPC for now and intend to use the unRAID box for less performance oriented things until I get a better feel for it. That is fine, just wanted to make you aware of that fact. The more drives you add the slower the parity checks will be but for general use you should not really see any problem. Tomorrow night I'd like to put the ECS board through its paces and see if I can get that "odd" gigabit NIC to work. Post back on this one as well. I would love to see if that one ends up working. You may need to disable the realtek NIC in BIOS first to force unRAID to use the "odd" NIC. Regarding the Vantec card, it has 6 ports and is marketed as such but really only uses the SiI 3114 chip which is a 4 channel card. There's a jumper on the card to activate all 4 internal or 2 internal, 2 external. That is kind of cool. I did not read up much on it; just saw it was PCI and then commented on it. While longer term testing would be required to be conclusive, it appears that the Promise internal ports have no problem running my external enclosure even though I've seen reports on newegg of this being flaky with this card. I think I'm keeping it. While the SATA II spec may not really make any noticeable difference over say a SiI 3114 based card, I like the fact that it supports NCQ and the like. So now that I have it up and running, what do you suggest I try and to to ensure it will be a robust, platform? Set up some drives, move some data (very large single files and very large numbers of small files) to it, check md5 checksums, run parity builds, and parity check, force a drive to "disable" itself and test out a rebuild of the data on it. That should get you started. Your report of the NIC being recognized looks good. unRAID loaded the sis900 driver Mar 2 22:02:09 Slipstream ifplugd(eth0)[1021]: Using interface eth0/00:0B:6A:DC:A6:CA with driver <sis900> (version: v1.08.10 Apr. 2 2006) and your board continued setup from there. Just a quick question about the name of your server; Slipstream. You are not by chance an Andromeda fan are you? The only reason I ask is because I have been going back through and watching all of the episodes recently and the name of your tower reminded me of that.
March 5, 201016 yr Author Post back on this one as well. I would love to see if that one ends up working. You may need to disable the realtek NIC in BIOS first to force unRAID to use the "odd" NIC. Here goes nothing. Pulled down the latest build and started a new USB key. This board didn't have the option to boot from specific USB key like the ASRock did. After trying all the options (USB-FDD, USB-CDROM, USB-ZIP) I scoured the web for how to boot it from USB. On the ECS website I found out that hitting F11 would bring up a boot menu. After disabling the full screen boot logo, I was able to access the boot menu, choose HDD+ which allowed me to select my USB key. From this point on, the USB key was listed as the primary boot device in the BIOS. It appeared to boot properly but did not seem to get an IP address with the gigabit controller enabled (and 100mbit Realtek disabled). Sure enough, I got "Device not found" from the console. Time to disable the "odd" gigabit LAN and enable the Realtek based one. With the Realtek 100Mbit NIC enabled, it booted but //tower was still not accessible. This time, the console indicated that the device was detected and entering the IP address in worked. Onward to testing drives! - Drive plugged into ATI SB chipset with ports set to IDE Mode (This board has 4 sata ports off this chip). - Worked. - Same as above but in AHCI mode. - Bios blinked saying no arrays were defined but the drive showed up fine. - JMicron 363 chip in IDE mode (only other option is RAID) using esata enclosure. (This chipset is marketed as the ports to use for esata). - It appears to work as advertized. - Syba 2 port 3512 PCI card - worked. - Promise & Syba together did NOT work (from the start). The Syba bios loaded and quickly displayed some warning thereafter that I didn't quite read. The Promise BIOS didn't load. Promise card/drive didn't show. - Promise & JMicron together did NOT work (from the start). Drive hooked up to the Promise card didn't show. - Promise SATA300 TX4 alone - worked after I realized that it was hung up and not seated properly. I re-tested the above scenarios and they all played nicely together. Didn't bother trying the IDE ports. I believe one runs off the JMicron chip and the other... ATI? From there I tried combinations like: a drive plugged into the ati chipset, another on the JMicron and the Promise card (possible configuration later on). I've attached the syslog of this config. Keep in mind up until this point I haven't done extensive testing like data verification - just checking to see that the drives show up and can be initialized. Overall I'm happy it works because like you said, this will eventually make a better board with 6 sata ports onboard and 2x PCIe x16 slots (if I go with a PCI graphics card / headless) plus the x1 slot for better controllers. The other 2 PCI slots could come in handy too. I'm a little sad there's no (out of box) support for the gigabit NIC as that would obviously be more useful and not require the added cost / slot for a separate one. I can look into support for that when the time comes, I guess. Alright, time to get my hardware back to the ASRock board and start doing some real stability testing there. Just a quick question about the name of your server; Slipstream. You are not by chance an Andromeda fan are you? The only reason I ask is because I have been going back through and watching all of the episodes recently and the name of your tower reminded me of that. You nailed it That's exactly the inspiration for the name. Seemed fitting as well since I plan to stream my media from this server. That and all my computers' names start with S for some reason. Is it worth re-watching? I used to catch episodes here and there in no particular order. I had been considering going back to the beginning... syslog-2010-03-04.txt
March 5, 201016 yr Here goes nothing. Pulled down the latest build and started a new USB key. This board didn't have the option to boot from specific USB key like the ASRock did. After trying all the options (USB-FDD, USB-CDROM, USB-ZIP) I scoured the web for how to boot it from USB. On the ECS website I found out that hitting F11 would bring up a boot menu. After disabling the full screen boot logo, I was able to access the boot menu, choose HDD+ which allowed me to select my USB key. From this point on, the USB key was listed as the primary boot device in the BIOS. Glad to hear you got it booted from the flash drive after a little bit of messing with it. It appeared to boot properly but did not seem to get an IP address with the gigabit controller enabled (and 100mbit Realtek disabled). Sure enough, I got "Device not found" from the console. Time to disable the "odd" gigabit LAN and enable the Realtek based one. With the Realtek 100Mbit NIC enabled, it booted but //tower was still not accessible. This time, the console indicated that the device was detected and entering the IP address in worked. Onward to testing drives! Good to hear that at least one of them is working. I was hoping that the "odd" NIC would work, but we were not quite that lucky. Overall I'm happy it works because like you said, this will eventually make a better board with 6 sata ports onboard and 2x PCIe x16 slots (if I go with a PCI graphics card / headless) plus the x1 slot for better controllers. The other 2 PCI slots could come in handy too. I'm a little sad there's no (out of box) support for the gigabit NIC as that would obviously be more useful and not require the added cost / slot for a separate one. I can look into support for that when the time comes, I guess. For the NIC problem I would send an email off to limetech with the board that you are using and the NIC chipset that is used, if you can find it. Usually you can visit the manufacturers web site and look through the driver downloads page to get a clue as to what chipset is used for what components. Once limetech knows about it he should be able to find the linux driver for that NIC and add it into the unRAID kernel. You nailed it That's exactly the inspiration for the name. Seemed fitting as well since I plan to stream my media from this server. That and all my computers' names start with S for some reason. Is it worth re-watching? I used to catch episodes here and there in no particular order. I had been considering going back to the beginning... Score one for me! I feel a like a geek knowing that... but if the shoe fits... I enjoyed rewatching all of the episodes and so did my room mate. I had seen them all before but he had missed random ones here and there and enjoyed getting the entire picture. We are moving on to Stargate Atlantis next, then maybe Stargate SG-1.
March 5, 201016 yr Author I seem to be having issues. The drives I am using were formerly in a Raid 0 array. When I was testing unRaid on the ECS board the controller reminded me that while the array was no longer there, the drives were still configured for an array so I deleted those config files. Bringing the drives back to the ASRock board didn't go so well. Last night I tried and tried to get the shares to show up but for some reason I would hit start, then if I refreshed, the drives would be stopped again. I tried updating to 4.5.3 and nothing changed. Randomly, I decided to format the drives back to NTFS and re-start the process. This time introducing them to the system indicated they needed to be formatted so I'd do that and it would indicate "formatting" but again, a refresh would make it look like nothing's happened. In the console I see stuff like line 1: 1263 Aborted nkreiserfs -q /dev/n d2 >/dev/null 2>1 and something similar for the 2nd drive. Refreshing also uncovered one error per drive. Edit: I grabbed the preclear script from the forum and ran that on both of my drives then re-introduced them and all seems fine now. Later, I would like to post screenshots of the results of the preclear though and perhaps get some help interpreting them to see if there's any cause for concern. Coming from the Windows world, Unraid could use better indications of what exactly is going on. Prostuff, I'll take your advice and suggest support for the LAN chip be added. Score one for me! I feel a like a geek knowing that... but if the shoe fits... Haha you should have seen my gf's face when I told her the origin of the name. What can I say? Then again she's no stranger to sci-fi either and I believe she called her eeepc Stratosphere... No matter what shows we go back and re-watch, none of them are going to fill the void BSG left. Not even Caprica. That might be a topic for another forum lol.
March 15, 201016 yr Author So I decided to just start using the box last night. I copied ~80GB of data to it and ran a parity check before bed. This morning, I was greeted with almost 5 million errors! Disk status was still all "green". What would cause this? Unfortunately I did not capture the syslog as I only had time to shut it down before heading out.
March 15, 201016 yr Run smart tests on the drives. Run memtest. Rerun parity test. Post and examine syslog.
March 15, 201016 yr Run smart tests on the drives. Run memtest. Rerun parity test. Post and examine syslog. I'd do the memory test first. Unless the voltage, timing, and clock speeds are exactly as specified by the manufacturer for the specific memory strips installed, then it is the most likely. Without a syslog we really can give any more than general advice. It is as if you said you went to drive your car to California but did not get there... What went wrong? No gas/ran out of gas... No battery or dead battery Bad spark-plugs (some cars still have them) Engine failed. Transmission failed. You never learned to drive. You learned to drive, but are male, refused to ask for directions (I think it IS a rule) and went eastward from Kansas. You learned to drive, but are female, did ask for directions (I think it IS a rule) and went eastward from Kansas. (no compass) Car is crushed (You were watching a Godzilla Movie and got carried away) Bridge/Highway is out (You were watching 2012) Tires missing, car on blocks (you parked it overnight in a rough neighborhood) Engine missing (an even rougher neighborhood) Car missing (Stolen... The worst kind of neighborhood) California missing (We had another huge earthquake/You were watching 2012) Without a trip log (syslog) we have no clues. Yes, most of the above is in fun. Odds are you just were in a hurry this morning. But still, a memory test is first priority.
March 16, 201016 yr Author LoL! Cali would be some drive for me. Likely my old wheels would fall off before I got there. General advice was all I could ask for this morning. I woke up to nearly 5 million errors and didn't even have time to get rid of weekend scruff so I just shut everything down and ran out the door. So let's see what steps I can re-trace prior to this event. - I had just moved my server from my work bench to the shelf underneath the staircase where it will live. - I had tried swapping out my longer eSata cable to make my external enclosure (housing disk 1 at the time) work in the new space - After realizing new cable was not appreciated by the promise controller, I swapped back out to the original that had worked and made the physical arrangement work. - I transferred a bunch of data to it (sorting as I go, so mostly little chunks at a time) with the exception of a ~22GB move of data just before going to bed. - Everything looked good, no errors reported so I started the parity check. We know the rest. My first suspect was the whole promise controller > esata bracket > esata cable > enclosure arrangement so I dug around for another 4-pin esata power adapter and plugged disk 1 in direct as the rest are. I've read quite a bit around the forums that cables can be an issue and I figured that could fall into that category. I then started Memtest (love that it's a built in option by the way) and just over an hour into it I see no errors. I know I should really let it run for a good amount of time (especially since I haven't done so yet with this setup) but I think I'd rather start another parity check tonight to see if brining the drive internal helped at all. I'll try to wake up tomorrow morning in time to de-scruffify AND at least collect syslogs if they haven't improved. If they haven't, I might consider that road-trip South after all... Edit: It appears my suspicions were correct as this morning's check led to 0 errors. I knew that the whole external setup was rather fidgety and this just confirmed it. Now I just need to get more creative with finding ways to get proper ventilation with 3 drives in that tiny case. (Hope tax season blesses me with new case money )
March 16, 201016 yr Remember that parity checks update parity as they go. So just because parity is now consistant with the data disk content, it doesn't mean the data didn't get scrambled as it was written. I'd do some serious checking on each data disk to make sure your data is correct.
Archived
This topic is now archived and is closed to further replies.