June 11, 200917 yr Hello forum members, I have an unRAID Plus server which I built about 3 weeks ago. It has one party drive (1Tb) and 4 data drives (2x 1Tb and 2x 750Gb) and has been running fine since I put in service. Tonight while ripping a CD to the server, I got Write errors from the Eaxct Audio Application. I tried to connect to the server via Windows Explorer but got an error message saying the request timed out. Next i went and looked at the server and the HD light was on constantly. So i powered off the server, connected a monitor and keyboard and re-booted the server. The bzimage and bzroot load fine. But here is where the problems starts. First, the keyboard locks up and I am unable to type any commands at the command prompt! Once after trying this a few times, I was able to type ifconfig and then the keyboard locked up! So, I can't reach the server via Web interface or type any commands at the terminal. What can i do? The way it just died suggests it may be a hardware failure of some sort, but because i can't get to unRAID to diagnose it, I am stuck in the water! If I have to replace the motherbaord and everything on it i will, but I'm not sure unRAID will rebuild" my array. Please help! I have all my movies and music on the server. Thanks in advance.
June 11, 200917 yr Author ...OK an update. I took a gamble and assumed the the drive to where i was ripping the CD might be defective as that was the drive that was in use when the failure occurred. So I disconnected it and re-booted the unRAID server. Voila! The server booted and i was able to access it via the web interface. The drive where my music is stored is the only PATA drive in the array and I (for shame) had it connected to the motherboard via a ROUND IDE cable. I remember reading somewhere that round IDE cables may have impedance/timing issues, so I replaced it with a flat cable and re-booted. The server mounted all disks and then proceeded to do a parity check. At the current rate it looks like it will take 5 hours to complete. So now I have three questions: 1. If the problem does indeed turn out to be a round IDE cable, is this what i can expect to happen? The server to lockup without warning? 2. It may turn out to be the disc is bad. If so, I will need to replace it and since PATA drives are like hen's teeth and I don't have any more available SATA ports, can i use a SATA drive with one of those PATA to SATA converters? Or should i get a PCI card with SATA ports? 3. Since the failure occurred while writing a specific file, can i cancel the parity check and and just erase the file fragment? Thanks!
June 11, 200917 yr 2. It may turn out to be the disc is bad. If so, I will need to replace it and since PATA drives are like hen's teeth and I don't have any more available SATA ports, can i use a SATA drive with one of those PATA to SATA converters? Or should i get a PCI card with SATA ports? i use a sata to pata (ide) converter without any problems. i also use sata pci cards without problems.
June 11, 200917 yr Hello forum members, I have an unRAID Plus server which I built about 3 weeks ago. It has one party drive (1Tb) and 4 data drives (2x 1Tb and 2x 750Gb) and has been running fine since I put in service. Tonight while ripping a CD to the server, I got Write errors from the Eaxct Audio Application. I tried to connect to the server via Windows Explorer but got an error message saying the request timed out. Next i went and looked at the server and the HD light was on constantly. So i powered off the server, connected a monitor and keyboard and re-booted the server. The bzimage and bzroot load fine. But here is where the problems starts. First, the keyboard locks up and I am unable to type any commands at the command prompt! Once after trying this a few times, I was able to type ifconfig and then the keyboard locked up! This is not usually a symptom of a disk drive problem, especially if you just rebooted. So, I can't reach the server via Web interface or type any commands at the terminal. What can i do? The way it just died suggests it may be a hardware failure of some sort, but because i can't get to unRAID to diagnose it, I am stuck in the water! If I have to replace the motherbaord and everything on it i will, but I'm not sure unRAID will rebuild" my array. Even if you had to replace the motherboard, your data on the disks would still be fine. If you do switch hardware, just be certain to assign the parity drive on the devices page to the same physical disk it was on the old motherboard. (It is OK if it is on a different disk-controller, and odds are it will be) Do a screen-print of the "devices" pages in the management interface NOW... while you remember, so you will have it later if you need it. ..OK an update. I took a gamble and assumed the the drive to where i was ripping the CD might be defective as that was the drive that was in use when the failure occurred. So I disconnected it and re-booted the unRAID server. When "writing" to the array two disks are being written to... the data disk AND the parity disk. Either could have caused errors if they were defective. Only way to know what was happening would have been to capture a copy of the syslog for analysis. Too late now to get a record of the original error, since the syslog is in memory and the old contents gone since you rebooted. Voila! The server booted and i was able to access it via the web interface. The drive where my music is stored is the only PATA drive in the array and I (for shame) had it connected to the motherboard via a ROUND IDE cable. I remember reading somewhere that round IDE cables may have impedance/timing issues, so I replaced it with a flat cable and re-booted. The server mounted all disks and then proceeded to do a parity check. That is a good sign... it could be the cable did not properly communicate a command to the drive, or, it could be something else... memory voltage, timing, clock speed are PRIME candidates to be checked against the values needed for your specific RAM strips. Do NOT rely upon the MB BIOS to set it properly. Some do, many do not. At the current rate it looks like it will take 5 hours to complete. So now I have three questions: 1. If the problem does indeed turn out to be a round IDE cable, is this what i can expect to happen? The server to lockup without warning? I'd expect disk errors. They would show up in the syslog. Eventually, if the syslog grows to where it uses up all memory the system would start killing processes trying to reclaim some RAM. That would cause it to eventually kill off the unRAID processes that deliver the network shares and web-interface as it runs out of processes to kill that are running. I'd leave the round cable out... it would not help things, that's for sure. 2. It may turn out to be the disc is bad. If so, I will need to replace it and since PATA drives are like hen's teeth and I don't have any more available SATA ports, can i use a SATA drive with one of those PATA to SATA converters? Or should i get a PCI card with SATA ports? A smartctl report should be run on the disk... It will tell you the health of the drive. Unless you hear mechanical screeching noises when the drive spins I doubt if you can guess the relative health of any drive. Run a "smartctl" report on each drive and post the results... (Or, at a minimum, the drive you suspect) 3. Since the failure occurred while writing a specific file, can i cancel the parity check and and just erase the file fragment? Sorry... but it does not work that way. Let the parity check complete. Please help! I have all my movies and music on the server. Thanks in advance. you need to help us first. 1. Verify memory timings, voltage, and clock speed are correct for your RAM. Run a memory test overnight, or at least for a few full cycles. 2. Post a syslog. 3. Run smartctl on each of your drives and post the output. Unless you have an inexpensive source of IDE 750Gig drives, I'd use one of the PATA to SATA converters if you do not have an empty slot in your MB, or a $15 to $20 dollar SATA card. I've used both SATA to PATA converters to use a old PATA drive on a new SATA controller, and a PATA to SATA converter to use an SATA drive on a old IDE controller. Both work just fine. They cost almost as much as a SATA controller card, so if you need more than one, it might be better to get the controller card instead of the converter. Joe L.
June 11, 200917 yr Author Thanks Jupilerman and Joe. The parity check completed with 156 sync errors (corrected I presume as everything appears to be working now. And the server seems faster). How do I run a smartctl report? When i tried the command line shown in the wiki I got some error. There was some explanation in the wiki that said unRAID v4.4 or greater needed something else to obtain a smartctl report. What do I need to do? Thanks!
June 11, 200917 yr Please help! I have all my movies and music on the server. Sad Thanks in advance. you need to help us first. 1. Verify memory timings, voltage, and clock speed are correct for your RAM. Run a memory test overnight, or at least for a few full cycles. Did you do this yet? you did not say anything about the results? 2. Post a syslog. Last few versions of unRAID have the missing library needed to run smartctl. But, we cannot tell what version of unRAID you are running because you did not tell us, and you have not posted a syslog (which would tell us) 3. Run smartctl on each of your drives and post the output. Rather than guess what error message you saw, can you let us know? That way, we don't need to be as clairvoyant. ;D If it is complaining about a missing library, then either upgrade to the current version (most recent beta) of unRAID that has the library, or install the library. You need to type the following command to test your IDE based 750gig drive: smartctl -a -d ata /dev/hd? where the /dev/hd? will probably be /dev/hda for the first IDE drive in your system. However, depending on the cabling, it could as easliy be /dev/hdb, or /dev/hdc, or /dev/hdd. We can't guide you because you did not post a syslog. (you seeing a pattern here?). The wiki has the following text...Note: if you get an error like "error while loading shared libraries: libstdc++.so.6", then you are using a version (such as v4.4.2) that is missing a required library. Please see this post: http://lime-technology.com/forum/index.php?topic=2817.msg23548#msg23548 Did you click on the link to get to the post that describes how to install the missing library? did you read it? did you download the file? Do you need additional guidance on installing it? A few more details would help us a lot and let us know where you are stuck. It is good the parity check completed... (the server would be no faster, unless errors caused by the round cable were slowing things down as the disk retried again and again to write to the disk) Joe L.
June 11, 200917 yr Author 1. Timings, voltages and clock speed are all fine. 2. Server is DEFINITELY faster and more responsive and I can only conclude it had something to do with the round cable as that was the only thing I changed. 3. My unRAID version is 4.4.2. It is unRAID Server Plus. 4. I am a UNIX newbie and glazed over after reading how to install the required library AND the fact that I would have to re-install it if I reboot. I would much prefer someone tell me how to upgrade to the latest version that has the required library Thanks for your patience.
June 11, 200917 yr 4. I am a UNIX newbie and glazed over after reading how to install the required library AND the fact that I would have to re-install it if I reboot. I would much prefer someone tell me how to upgrade to the latest version that has the required library Thanks for your patience. Read the instructions in the Announcements forum on how to upgrade. It is only a matter of replacing only a couple of files and rebooting.
Archived
This topic is now archived and is closed to further replies.