June 14, 200917 yr I have been searching for an answer to this but can't seem to find any clear answer. I have unRaid setup on a small server with 3 drives attached. It works fine, can read and write with no problem EXCEPT when transferring either large files or a lot of files at one time. Every time we go to do this, it will hang/crash the server and it has to be re-booted. What do you think is going on? If you need some specific info from me, let me know. Appreciate any help, K
June 14, 200917 yr To help you, we need to see your syslog, captured as late as possible before it crashes. Please see my sig for the Troubleshooting link, for help with capturing a syslog. Also, I recommend tailing the syslog on the physical console of your unRAID server, see Viewing the System Log. Perhaps there will be a few last messages on the screen, just before it crashes. You can use a digital camera to 'image' the screen.
June 15, 200917 yr Author This is my current log. I will try and transfer a large file and grab a syslog before it resets.
June 15, 200917 yr Author Honestly... the day I come here to find a solution the server goes ahead and works. I have been trying to transfer this same file (about 4.5 gigs) to the server off and on for over a month. It would always hang up and crash at different intervals of transfer until now. shesh! Anyway, here is the successful transfer syslog. I'm not sure if it will show anything. If not, maybe I can try and make it crash by transferring multiple files. The wife was transferring a bunch of pictures over and it crashed on her yesterday.
June 16, 200917 yr Thank you for the syslog. Would you also provide some detail about your hardware setup? You appear to have a board with VIA chipsets, which unfortunately have had some compatibility issues, not that that is necessarily the problem, but could be a possibility. I recommend upgrading to the latest version of unRAID, v4.5-beta6, so that you can take advantage of whatever compatibility fixes have been added for VIA chipsets. See the Release Notes page for the download link. I would also recommend running an overnight memory test, the one on the unRAID boot menu, just so we can eliminate memory issues. The syslog looks fairly good, except for a strange problem at Jun 13 19:58:36. A parity check had finished at 17:59:41 (with very good speed!), then the 3 drives had spun down at 18:31:26, then something serious happened at 19:58:36, to all 3 drives. All 3 drives reported a PHY change, which I understand is similar to a hotplug (dis)connect, then were slow to respond, then all 3 drives returned to service, with no further issues. The only thing I can think of is, all 3 drives were jolted loose briefly, perhaps in a backplane, or there was an electrical spike, causing a brief loss of communications with the drives. Were you aware of anything at almost 8pm on the 13th, that could explain this? You are booting with the pci=nomsi boot option. Do you know why that particular option is there? What happens when you boot without it?
June 16, 200917 yr Your second syslog includes evidence of a crash just now, apparently involving the VIA network chipset. Jun 15 23:46:59 Tower kernel: ------------[ cut here ]------------ Jun 15 23:46:59 Tower kernel: WARNING: at net/sched/sch_generic.c:219 dev_watchdog+0xf0/0x16d() Jun 15 23:46:59 Tower kernel: NETDEV WATCHDOG: eth0 (via-rhine): transmit timed out Jun 15 23:46:59 Tower kernel: Modules linked in: md_mod via_rhine via82cxxx ide_core ahci libata Jun 15 23:46:59 Tower kernel: Pid: 0, comm: swapper Not tainted 2.6.27.7-unRAID #3 Jun 15 23:46:59 Tower kernel: [<c011cc08>] warn_slowpath+0x61/0x86 Jun 15 23:46:59 Tower kernel: [<f883ddd2>] ata_build_rw_tf+0x184/0x241 [libata] ... Jun 15 23:46:59 Tower kernel: [<c01019e1>] cpu_idle+0xbd/0xd5 Jun 15 23:46:59 Tower kernel: ======================= Jun 15 23:46:59 Tower kernel: ---[ end trace 7acb02bcfc39758d ]--- Jun 15 23:46:59 Tower kernel: eth0: Transmit timed out, status 0003, PHY status 786d, resetting... Jun 15 23:46:59 Tower kernel: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1 Jun 15 23:47:03 Tower kernel: eth0: Transmit timed out, status 0003, PHY status 786d, resetting... Jun 15 23:47:03 Tower kernel: eth0: link up, 100Mbps, full-duplex, lpa 0x45E1 It did seem to recover here, but I wonder if at other times, it was not able to recover. I definitely do not trust it. I'm not sure I can recommend that VIA-based motherboard, but if you wish to continue with it, I recommend purchasing a network card, to use instead of the onboard networking. It's an opportunity to obtain a gigabit NIC too. Edit: I should add that upgrading to the latest unRAID release *might* include later drivers and modules for the VIA chipsets, and that might include fixes for bugs that are affecting you, causing your crashes. It's worth a test.
June 16, 200917 yr I should add that upgrading to the latest unRAID release *might* include later drivers and modules for the VIA chipsets, and that might include fixes for bugs that are affecting you, causing your crashes. It's worth a test. [note repeated here so you would be sure to see it]
June 16, 200917 yr Author Hey Rob, Thank you for looking into this for me. I appreciate your time in this matter. I was trying to stay cheap with parts I had laying around the house in building this. Hence the cheap-o MB and no dedicated NIC. But if you feel they may be causing my issues I can look into changing them. Is it difficult to setup unRaid again if I change them out? I do not know exactly how the OS works. I know in Windows changing out hardware like this can cause issues, especially the MB. As far as an electrical spike of any sort, our electric grid in our neighborhood is actually pretty stable. In the three years I have lived in my home our power has never gone out. (Even during Hurricane Ike) I don't recall anything happening at all during that time. To me, it feels like something is getting overloaded or overworked. Because the server will hang whenever we try to rapidly access files from it too. Like saying I'm browsing through multiple mp3's rapidly it will do the same thing. Anyway... I'll see what the wife thinks about me purchasing a new MB and NIC. heh... this should go over well.
June 16, 200917 yr Author Ah, btw, sorry forgot to comment about this: pci=nomsi boot option I have to place that in for the system to work. Before I used it, unRaid would boot up but it would not recognize my hard drives. After I placed that command in, it worked great.
June 16, 200917 yr The basic question is, is it usable, reliable? If you can't trust your file server, what good is it? I'm all for cheap, have no money myself, but trustworthy usability is a basic requirement. You do have 3 more steps before giving up on this board: * Memory test (very easy, run overnight). If a memory test fails, then you can buy good replacement memory, and that should fix all or part of the problem with the board. * Upgrade unRAID (very easy and free). It is still possible that a newer version will have improved components for your board. Just extract bzroot, bzimage, and memtest to your flash drive and reboot, it's that easy. * Install cheap network card. It's possible to find old PCI network cards for $3 to $10, and that will allow you to test if the onboard NIC was the primary problem. A gigabit card would be better, but it will usually run about $20, and for about $30 more, you could get a better motherboard, with gigabit built in. Then you would probably need 1GB or 2GB of RAM for $25, and a CPU for $30 to $60. Moving to a new system is easy with unRAID. Just record the drive assignments on the Devices tab, change out the hardware, and check and correct the drive assignments in the new system.
June 17, 200917 yr This may be a tad off topic, but I was having the same problems you were. If I tried to copy two different files to the server at the same time it would lock up. If I tried to que multiple files to copy overnight it would lock up. I have several posts talking about this behavior. I tried every version of unRAID from 4.3 to 4.5b6, nothing worked. I couldn't keep the monster even running more than a day or two. Forget about putting the drives to sleep, they would never wake back up & we would lock up again. I don't have the same board as you but I was running a AMD nforce2 motherboard. I never had data corruption problems like some people had with this board. I just couldn't keep it running. I had a extra Intel P4 motherboard laying around & decided to try that. Wished I had done that a year ago instead of messing around. Drives wake up properly & I can run this machine for a month & it never locks up. It eats multiple file transfers for breakfast, never complains. I have the two motherboards listed under my username. I will never use a problem board again..
June 17, 200917 yr Similar story here, and my 2 boards are also listed under my name. You know what is worse than a board that doesn't work? A board that does work, MUCH of the time! Especially if you pride yourself on fixing things, and hate to give up on a project. I wasted months on my nForce 4 board, and I regret most of it. I don't like to ever tell others what to do in their situation, but I too will not waste any more time on a problem board.
June 18, 200917 yr Author Well then! New MB it is. Thanks again guys for responding. And I appreciate the simple how-to on re-setting up unRaid.
June 20, 200917 yr Please let us know when you fix your problem. Also tell us what you did to make your server more stable. It may help someone in the future..
Archived
This topic is now archived and is closed to further replies.