October 31, 200916 yr First post here, but I've been lurking for few months. I put together the following build: Case: Norco 4220 Mobo: Asrock M3A785GXH CPU: AMD Athlon II X2 3.0Ghz Memory: 2x2GB G.Skill 10666 PSU: Corsair 750TX Parity Drive: 2TB Hitachi HD32000 7200RPM USB Key: 2GB ATP EarthDrive I put the hardware together without problem, prepped the USB Key, loaded unRaid, and booted up first try without any problems. I was very impressed! Once I had everything running, I found out I had used the wrong GUID when ordering my key. I sent an email to support with the correct GUID and got a new key within hours. Impressed again! The array has been built and parity check found 0 errors. I attached my syslog to this post. Now for my questions: 1) The mobo has a built in Realtek 8111DL gigabit ethernet, but unRaid is only detecting a 100Mbit connection. I've swapped cat5e cables to make sure that wasn't the problem. Does this just leave me with a driver issue? Or am I missing something stupid? I know 8111B/C has driver support, so I may just be stuck waiting for a future release. 2) I have a blank 1.5TB and a 1TB drive installed so far as data drives. I also have another 1TB drive in my HTPC that is NTFS formatted and loaded to the brim with media. I want to migrate this drive over to the server, but what is the best way to transfer the data? Do I need to just transfer across the LAN? If I can't fix question 1, then this would be a very slow process. Or is there a way to integrate the drive with full data into the array? 3) Next question isn't so much about unRaid, but about the Norco 4220, which is fairly popular around here. I swapped out the stock fans with four quieter 30 CFM fans. This works alright with the 3 drives I have in there now (soon to be 4), but I don't think this will be adequate as I start stuffing it with more drives (I'd like to eventually have 16-20). I've read that a few users have modified the fan plate to use 120mm-140mm fans, but I'm not really comfortable modifying my only fan plate. If anyone has a 4220 with 15+ drives can comment as to what what their cooling solution is, I'd appreciate it. Are the stock fans overkill for a denser drive arrangement, or are they barely adequate (i.e. would I be able to get by with quieter, lower CFM 80mm fans as I add more drives)? Is the 120mm fan mod really the only other option? Thats it for my questions for now, but I'm sure I'll have more as I go along. I'll try to run the machine through the recommended tests once I get more settled in with it, so the mobo can get added to the compatible hardware list. Thanks!
October 31, 200916 yr First post here, but I've been lurking for few months. I put together the following build: Case: Norco 4220 Mobo: Asrock M3A785GXH CPU: AMD Athlon II X2 3.0Ghz Memory: 2x2GB G.Skill 10666 PSU: Corsair 750TX Parity Drive: 2TB Hitachi HD32000 7200RPM USB Key: 2GB ATP EarthDrive If you could take a look at the Hardware Compatibility Page and the Motherboard Rating System page that would be a big help. It looks like you are the first to use this board and I would love to get it added so other know it works. Just take a look at some of the "Tested Level" links from the Hardware Compatibility Page. I put the hardware together without problem, prepped the USB Key, loaded unRaid, and booted up first try without any problems. I was very impressed! Once I had everything running, I found out I had used the wrong GUID when ordering my key. I sent an email to support with the correct GUID and got a new key within hours. Impressed again! The array has been built and parity check found 0 errors. I attached my syslog to this post. Now for my questions: Glad to hear you got everything up and running with not to many problems. 1) The mobo has a built in Realtek 8111DL gigabit ethernet, but unRaid is only detecting a 100Mbit connection. I've swapped cat5e cables to make sure that wasn't the problem. Does this just leave me with a driver issue? Or am I missing something stupid? I know 8111B/C has driver support, so I may just be stuck waiting for a future release. My suggestion would be to try and force the 1GB connection from the command line of the unRAID server. if that does not work then maybe the router is at fault? If all that fails they and contact Tom (limetech) and see if there is an updated drive that can be included with the next beta (OR if your not on 4.5beta7 then try that one first). 2) I have a blank 1.5TB and a 1TB drive installed so far as data drives. I also have another 1TB drive in my HTPC that is NTFS formatted and loaded to the brim with media. I want to migrate this drive over to the server, but what is the best way to transfer the data? Do I need to just transfer across the LAN? If I can't fix question 1, then this would be a very slow process. Or is there a way to integrate the drive with full data into the array? Using JoeL's unMenu package you can put the drive into the machine and mount it and copy the data that way. I personally would just transfer it over the network and then move the drive to the unRAID machine and format it. If something were to fubar the unRAID machine while the drive is in there and a copy is happening then you would not loose the data on the drive since it was in the HTPC. 3) Next question isn't so much about unRaid, but about the Norco 4220, which is fairly popular around here. I swapped out the stock fans with four quieter 30 CFM fans. This works alright with the 3 drives I have in there now (soon to be 4), but I don't think this will be adequate as I start stuffing it with more drives (I'd like to eventually have 16-20). I've read that a few users have modified the fan plate to use 120mm-140mm fans, but I'm not really comfortable modifying my only fan plate. If anyone has a 4220 with 15+ drives can comment as to what what their cooling solution is, I'd appreciate it. Are the stock fans overkill for a denser drive arrangement, or are they barely adequate (i.e. would I be able to get by with quieter, lower CFM 80mm fans as I add more drives)? Is the 120mm fan mod really the only other option? There is a thread here somewhere (do some searching in the forum) so people that have messed with the fan plate in there Norco Cases. That should get you going in the right direction. Thats it for my questions for now, but I'm sure I'll have more as I go along. I'll try to run the machine through the recommended tests once I get more settled in with it, so the mobo can get added to the compatible hardware list. Thanks! If you have any more questions feel free to ask and welcome to the unRAID community.
October 31, 200916 yr Author My suggestion would be to try and force the 1GB connection from the command line of the unRAID server. if that does not work then maybe the router is at fault? If all that fails they and contact Tom (limetech) and see if there is an updated drive that can be included with the next beta (OR if your not on 4.5beta7 then try that one first). How do you force a 1GB connection? "ethtool -s eth0 speed 1000"? I tried that and no luck. I checked and I'm running 4.5 beta6, so I'll try out beta7 and see if that helps.
October 31, 200916 yr My suggestion would be to try and force the 1GB connection from the command line of the unRAID server. if that does not work then maybe the router is at fault? If all that fails they and contact Tom (limetech) and see if there is an updated drive that can be included with the next beta (OR if your not on 4.5beta7 then try that one first). How do you force a 1GB connection? "ethtool -s eth0 speed 1000"? I tried that and no luck. I checked and I'm running 4.5 beta6, so I'll try out beta7 and see if that helps. What do you see when you type: ethtool eth0 It should show you what is available and negotiated for your interface chipset. Output will look like this: root@Tower:/# ethtool eth0 Settings for eth0: Supported ports: [ TP ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Advertised auto-negotiation: Yes Speed: 1000Mb/s Duplex: Full Port: Twisted Pair PHYAD: 0 Transceiver: internal Auto-negotiation: on Supports Wake-on: umbg Wake-on: g Current message level: 0x00000007 (7) Link detected: yes
October 31, 200916 yr Author Settings for eth0: Supported ports: [ TP MII ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Advertised auto-negotiation: Yes Speed: 100Mb/s Duplex: Full Port: MII PHYAD: 0 Transceiver: internal Auto-negotiation: on Supports Wake-on: pumbg Wake-on: d Current message level: 0x00000033 (51) Link detected: yes
October 31, 200916 yr Settings for eth0: Supported ports: [ TP MII ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Advertised auto-negotiation: Yes Speed: 100Mb/s Duplex: Full Port: MII PHYAD: 0 Transceiver: internal Auto-negotiation: on Supports Wake-on: pumbg Wake-on: d Current message level: 0x00000033 (51) Link detected: yes It would appear to me as if whatever you are connecting to, or the cable you are using is unable to support 1000Mb speed. The chipset driver is advertising it, but the auto-negotiation is unable to support that speed. Did you make the cable yourself? or is it pre-made. Does it have all 8 conductors connected at both ends? Is it rated for cat5e, or Cat-6? Older cables can't support the higher speed. What is the other end of the cable plugged into? Does your switch/router support 1000MB/s speed? Joe L.
October 31, 200916 yr Author It would appear to me as if whatever you are connecting to, or the cable you are using is unable to support 1000Mb speed. The chipset driver is advertising it, but the auto-negotiation is unable to support that speed. Did you make the cable yourself? or is it pre-made. Does it have all 8 conductors connected at both ends? Is it rated for cat5e, or Cat-6? Older cables can't support the higher speed. What is the other end of the cable plugged into? Does your switch/router support 1000MB/s speed? Joe L. Both my HTPC and the unRaid server are plugged into an HP ProCurve 1400 gigabit switch. The switch shows a 1000MB/s connection between it and the HTPC. I took the cable connecting the HTPC to the switch and swapped it out with the one connecting the switch to the server, with no change, and the switch still showed a gigabit connection to the HTPC. Both cables are pre-made, and both are rated for cat5e. I also double checked the BIOS to make sure there wasn't anything I overlooked, but didn't see anything applicable. Could it be a problem with the mobo/ethernet port? I'm hesitant to jump straight to that conclusion, especially since there is a stable connection, albeit at 100 and not 1000. Thanks for the suggestions so far.
November 1, 200916 yr Author My server won't boot anymore. I was still trying to fix the LAN speed issue, and tried to force a gigabit connection with the following command: ethtool -s eth0 speed 1000 duplex full autoneg off. This caused the server to hang, so I restarted. Once I restarted, the server wouldn't boot to the login prompt anymore, it would hang at the point where it outputted Tower kernel: sd 0:0:0:0: [sda] 3907029168 512-byte hardware sectors: (2.00 TB/1.81 TiB) (I couldn't get a syslog saved). I decided this would be a good time to upgrade to beta7 to see if it fixed my problem (by copying over the two bz files with the USB plugged into another machine). Doing this got the machine to boot into a message similar to this post, followed by a line similar to "no more processes to run at this level" and a 5 minute wait, at which point it would repeat. I then went back to beta6 again, and this time disabled UnMenu from loading. This got me to the login prompt, but I could not run ethtool or smartctl, and no other machines could detect the server on the LAN (I tried to copy the syslog to a text file at this prompt, but once I checked the USB key it wasn't there). I restarted once more, and this brought up the login prompt once more, followed by a string of random (to me at least) alphanumeric characters that would repeat every few seconds. Any thoughts?
November 1, 200916 yr My server won't boot anymore. I was still trying to fix the LAN speed issue, and tried to force a gigabit connection with the following command: ethtool -s eth0 speed 1000 duplex full autoneg off. This caused the server to hang, so I restarted. Once I restarted, the server wouldn't boot to the login prompt anymore, it would hang at the point where it outputted Tower kernel: sd 0:0:0:0: [sda] 3907029168 512-byte hardware sectors: (2.00 TB/1.81 TiB) (I couldn't get a syslog saved). I decided this would be a good time to upgrade to beta7 to see if it fixed my problem (by copying over the two bz files with the USB plugged into another machine). Doing this got the machine to boot into a message similar to this post, followed by a line similar to "no more processes to run at this level" and a 5 minute wait, at which point it would repeat. I then went back to beta6 again, and this time disabled UnMenu from loading. This got me to the login prompt, but I could not run ethtool or smartctl, and no other machines could detect the server on the LAN (I tried to copy the syslog to a text file at this prompt, but once I checked the USB key it wasn't there). I restarted once more, and this brought up the login prompt once more, followed by a string of random (to me at least) alphanumeric characters that would repeat every few seconds. Any thoughts? I would double-check the memory timing, voltage, and clock speed settings on the MB. Too many un-related problems all at once to not give them a check. Most motherboards attempt to set those parameters, but many get it wrong, especially with premium RAM. It must be correct for your specific memory strips. Might also want to run the memory test for a few cycles, just to be sure. Joe L.
November 1, 200916 yr Author I would double-check the memory timing, voltage, and clock speed settings on the MB. Too many un-related problems all at once to not give them a check. Most motherboards attempt to set those parameters, but many get it wrong, especially with premium RAM. It must be correct for your specific memory strips. Might also want to run the memory test for a few cycles, just to be sure. Joe L. You are quite correct. When I first put the system together, I ran memtest for 2 complete cycles with no errors. I ran it just now, and every single address is coming up bad (I had over 100,000 errors within 5 minutes of running the test. The test crashed shortly after). Why would every single address show up bad now, only 2 days after both modules showed 0 errors? Should I be concerned that the mobo is the problem, and not the memory (I switched the memory to different slots, but got the same results)? I am also concerned that these problems began to snowball after I tried forcing a gigabit connection in unRaid. Is there any way software caused this, or is it just a coincidence?
November 1, 200916 yr I would double-check the memory timing, voltage, and clock speed settings on the MB. Too many un-related problems all at once to not give them a check. Most motherboards attempt to set those parameters, but many get it wrong, especially with premium RAM. It must be correct for your specific memory strips. Might also want to run the memory test for a few cycles, just to be sure. Joe L. You are quite correct. When I first put the system together, I ran memtest for 2 complete cycles with no errors. I ran it just now, and every single address is coming up bad (I had over 100,000 errors within 5 minutes of running the test. The test crashed shortly after). Why would every single address show up bad now, only 2 days after both modules showed 0 errors? Should I be concerned that the mobo is the problem, and not the memory (I switched the memory to different slots, but got the same results)? I am also concerned that these problems began to snowball after I tried forcing a gigabit connection in unRaid. Is there any way software caused this, or is it just a coincidence? Probably just a coincidence. Software won't hurt the RAM. Did you double-check the BIOS settings? Joe L.
November 1, 200916 yr You are quite correct. When I first put the system together, I ran memtest for 2 complete cycles with no errors. I ran it just now, and every single address is coming up bad (I had over 100,000 errors within 5 minutes of running the test. The test crashed shortly after). Why would every single address show up bad now, only 2 days after both modules showed 0 errors? Should I be concerned that the mobo is the problem, and not the memory (I switched the memory to different slots, but got the same results)? I am also concerned that these problems began to snowball after I tried forcing a gigabit connection in unRaid. Is there any way software caused this, or is it just a coincidence? Did you check the memory timings as Joe suggested? That would be my first thought. Nothing you did in unRAID should have affected anything after a power cycle of the computer. If your memory timings are good, but the memory is failing right and left - either the memory got zapped or the motherboard is bad. If possible, I'd try the memory on another motherboard to see if it is okay. If it is, at least you know that the motherboard isn't damaging the memory. If it is bad in the new motherboard as well, RMA it and try another set.
November 1, 200916 yr Author I checked the BIOS, everything was set to AUTO. I put in manual settings (Voltage to 1.65V. Memory timings tCL 9 - tRCD 9 - tRP 9 - tRAS 24, according to manufacturer), and ran the test again. Same result. The 30 day timeframe for returning parts actually expired yesterday (being stationed in Japan means I'm waiting 2 weeks for Newegg shipments), so I may be stuck coughing up more money. I'll try to get a hold of some more RAM locally and see if that works, since I don't think I'll have access to another mobo to test with. Thanks for the responses so far.
November 1, 200916 yr Since you have 2 memory strips, you might try running with just one, or the just the other. It is very unlikely for both to have failed in exactly the same way at the same time. (Possible, but un-likely) Perhaps it is just one memory strip that has failed. If it fails the test with either one, then a try with a third strip obtained locally might point to the MB as the most suspect. (You did put a heat-sink on the CPU? with heat-sink-compound? ... You probably did, but I'm just trying to rule out a CPU overheat causing the problem. Is the heat-sink still attached and not come loose?)
November 1, 200916 yr Try increasing tRFC, should be around 195us or 52-55 with 2GB/1066 memory. A lot of early AM2+ / AM3 boards were setup for 1GB DIMMs not 2GB DIMMs and the SPD values arent pulled correctly.
November 1, 200916 yr Also check your USB is plugged back in the original slot, I had a weird issue with one motherboard where my network wouldnt connect with the usb key in a front port, worked great in a rear port, turned out to be an IRQ issue. Wont help with memory but might help with 1000Mb connection.
November 2, 200916 yr Author Since you have 2 memory strips, you might try running with just one, or the just the other. It is very unlikely for both to have failed in exactly the same way at the same time. (Possible, but un-likely) Perhaps it is just one memory strip that has failed. If it fails the test with either one, then a try with a third strip obtained locally might point to the MB as the most suspect. (You did put a heat-sink on the CPU? with heat-sink-compound? ... You probably did, but I'm just trying to rule out a CPU overheat causing the problem. Is the heat-sink still attached and not come loose?) You were right. I have it running now with one 2gig stick of ram, and it boots up fine once more. I am still stuck with an ethernet speed of 100Mb/s though. Also check your USB is plugged back in the original slot, I had a weird issue with one motherboard where my network wouldnt connect with the usb key in a front port, worked great in a rear port, turned out to be an IRQ issue. Wont help with memory but might help with 1000Mb connection. I double checked that I'm using the same USB port, and even tried switching to a couple of different ones. No change though. I did go back to trying the command ethtool -s eth0 speed 1000 duplex full autoneg off, and ethtool returned a speed of 1000! But with autoneg off, it wouldn't detect a link any more. I turned autneg back on, and speed dropped back down to 100, but a link was detected again. I included a new syslog, if it helps at all. UnMenu showed 4 orange lines in the log, but it didn't look like it was related to network stuff. Thanks again for the help so far!
November 2, 200916 yr Try moving your USBs keyboard to the same pair of USB ports as your USB key. Two of your USB ports appear to be using IRQ 18 same as your NIC. Normally these would be masked out to different IRQs via MSI/MSI-X I see NIC masked out to 22 but not the USB ports. Worth confirming in the bios everything that can be turned off or disabled is.
November 3, 200916 yr Your first syslog was perfect, everything looked good, and if you had not indicated otherwise, I would have assumed that you had a gigabit connection, because as Joe said, it announced itself and its capabilities as a gigabit driver and chipset, with no indication of having to negotiate a slower connection. It would not surprise me if it is occasionally negotiating a gigabit connection, and you may have seen one, briefly. Your second syslog shows half the RAM of course, but it also indicates considerable trouble keeping a network connection. It could not immediately connect, then did succeed for about 8 minutes, then went down for about 5 minutes, came back up for about a half a minute, was down briefly, then back up. Perhaps that was your own testing, unhooking and rehooking and/or recycling the switch/router? Unfortunately, the r8169 driver does not report the negotiated speed of each connection made, as do other network drivers. The ethtool eth0 command and the UnMENU Network plugin will show you your current network connection speed. See the Console Commands for Networking section for other useful network diagnostic commands. In particular, try the ifconfig command. It will show you if there are errors and collisions on your network connection (there should be ZERO errors and collisions). If you see errors, then the connection is bad, either a bad cable, bad chipset or connector in your server, or possibly power line interference near the network cable (don't run power lines right next to network lines).
Archived
This topic is now archived and is closed to further replies.