October 14, 20169 yr Hi Everyone - I am facing some strange issues on my server.2 of the last 3 nights, while I have been trying to preclear a new drive, I have come to check my server in the morning and found it to be hard crashed. After rebooting, everything comes up as normal, array starts up, all is peachy. Today, I decided to tail syslog into a file on /boot to see if it happened again if I could catch the issue, while at it, I invoked mover just to check the tail was working properly. When doing so my telnet session dropped (not sure if it relevant, but I was unable to SSH to the server today, which I have done for the past 6 weeks since building the server), I refreshed the "Main" page in GUI to see my Array was offline, but the Server was running. So, I restarted the Array, tried again, same action. While watching the log, nothing happened, it just froze while Mover was running. Only 3 TV shows, ~5 GB, were on cache at the time. Nothing got moved. I invoked mover manually from command line, and everything moved as expected. Hurray! So, I wanted to test a theory of an issue with the Mover in UI. I tried to copy 4GB back to the cache (mv file1 file2 file3 /mnt/cache/TV) and the array crashed again. Everything has been running smoothly (mover daily at 4am, always something on cache to move) for the last 4 weeks since I was all setup and configured properly, until 3 days ago when i started waking up to crashes. Diagnostics.zip are attached, system hardware details below. I really appreciate any advice. For now I am going to pull out the video card and see if that fixes anything (new crappy card I had to buy when doing initial setup so I could get into BIOS, no onboard video). System info - Motherbaord - BIOSTAR A770 A2+ AM2+/AM2 AMD 770 ATX AMD Motherboard CPU - AMD Phenom 9850 BLACK EDITION Quad-Core 2.5 GHz Socket AM2+ 125W HD985ZXAGHBOX Processor Graphics - EVGA 512-P3-1300-LR GeForce 8400 GS 512 MB DDR3 PCI Express 2.0 RAM - 8 GB Mushkin Enhanced DDR2 SDRAM PSU - Orion HP585D Parity - WD Red 3TB Array - WD Blue 2TB - Samsung 1TB - Seagate 1TB Cache - Samsung 840 EVO 250 GB Flash - Data Travler 2.0 1GB Unassigned - HGST 4TB PCI Device - IO Crest 4 Port SATA III PCI-e 2.0 x1 Controller Card Marvell Non-Raid https://www.amazon.com/gp/product/B00AZ9T3OU/ref=oh_aui_detailpage_o08_s00?ie=UTF8&psc=1 Also using this molex to SATA adatpers https://www.amazon.com/gp/product/B00009YFTI/ref=oh_aui_detailpage_o06_s00?ie=UTF8&psc=1 UnRaid Version 6.2 Dockers - Plex, OwnCloud, mySQL, No-IP, OpenVPN nas-diagnostics-20161014-1035.zip
October 14, 20169 yr Community Expert Is that drive you are trying to preclear plugged into that PCI-e SATA controller? If it is, try plugging that drive into a mother board SATA port and see if it locks up the server during preclear. (When I was researching SATA controllers, there was some indication that some of them had heat issues when active for a long period of time. Ones with REALLY large heat sinks seem to work better...)
October 14, 20169 yr Author No Floppy. Only drives are the hdds and ssd mentioned. Yea, the drive I am trying to pre clear is in the sata controller. I'll try moving it to the board. Thanks for the tip.
October 14, 20169 yr Author So bigger issue now.... Server crashed again, now I can't boot anymore. System takes very long time to post (2+ minutes), then after post says "Disk boot error, insert system disk and press enter". Tried moving the flash to different usb ports, checked bios and boot order (booting from USB-FDD, all else disabled). Even plugged it into my windows desktop, reformatted and tried again with a new download. I'm starting to lose it here. Any help would be great.
October 14, 20169 yr Community Expert Double check in the BIOS that you are still trying to boot from the Flash Drive. Occasionally, some BIOS will assume that any new disk is the boot disk. Second suggestion is to stick your flash drive into another computer and see if will boot there. (Don't worry, unRAID will not write to any drive unless it already has the drive's Serial Number in its configuration file.)
October 15, 20169 yr Author Thanks...not sure what happened, but my boot device became corrupted, so I had to reformat it. Bios boot prefer alps got reset somehow, but everything is booting now. Now back to trying to figure out why my array keeps crashing.
October 15, 20169 yr I've been having the same issue the last few days. I upgraded to 6.2.1... A few days after, I added a pcie sata card and 4 more drives. The system booted fine, all the drives cleared formatted without issue and I was able to assign them. I copied 400GB over to one without issue. A few days after that the system would just crash, completely unresponsive. I have to hit the power button to reboot. After coming back up, it would work for a few hours and crap out again. I tried a new sata card, but same result. I've removed the new drives and sata card to see if fixes it.
October 15, 20169 yr Author I have a pci data card as well that has been in use for over a month without issue. I tried to start a parity check last night and it caused my array to go offline again. Server is still up, but the array goes down. So the pattern seems to be whenever the system is working harder than usual. It has now gone offline during pre-clear, while running mover, and while doing parity check. My guess is some type of hardware issue at this point. I have also noticed that it is taking a very long time to post (2 minutes sometimes), and there is no "beep" at all during startup. Yesterday I pulled out the video card, pic data card and each stick of memory (1 by 1) and disconnects every hard drive...so it was basically booting from the flash drive with only the motherboard, cpu, psu connected. I'm afraid it's going to end up being motherboard related...
October 15, 20169 yr Community Expert One other possibility is the PS. Have a good look at its spec's. Does it have a dual or single 12V rail? Is its wattage level really there for the new load generated when you added the new drives and SATA card? Is a quality unit or a cheap one from a no-name manufacturer?
October 15, 20169 yr Author I would definitely prefer the issue to be the PS. It is definitely a cheaper off-brand PS. It is going on probably 5 years old...As is noticibale from my specs, this is an 8 year old PC build. After 2 years I upgraded and gave it to my parents, while they had it the original PS died, and I was in town for just the day to fix it (not knowing the issue when I got there) so ran up to Best Buy to grab another PS. So the PS in it is probably 5 years old, which for off-brand is probably a bit old. 1x 12V Rail 585 Watt; Should be plenty to power what Im running (All listed in my OP). Link to the New Egg Specs Page in case anyone wants to take a look and see if I am missing something. http://www.newegg.com/Product/Product.aspx?Item=N82E16817339012 I think I am going to order a Corsair or Seasonic (Have a Corsair in my main PC and it has ran it wonderfully for years).
October 15, 20169 yr Community Expert Look at the spec's on the side of this unit (attached) and use your calculator. Your CPU alone has 125W power envelope! And then there is the SATA card and the Graphics Card...
October 15, 20169 yr Author Wow...thanks for pointing that out...clearly I don't know much about PSUs...I have always went off of the Wattage that they claim (the one in there said 585w, so I plug it into a calculator with all my components and it said I needed 380, assumed I was good), never looked into the wattage breakdown across the different rails...Learn something new everyday Looking at the attached PSU now... The graphics card specs say it runs off of the 12V rail, requiring 350w. Cant find any power requirements on the SATA card. Hopefully this PSU is enough?
October 15, 20169 yr Personally, I wouldn't trust that supply, as the manufacturer is completely out to lunch #1 - As stated, the 12V line pumps out 46A which they state is 522W. Incorrect 46A is 552W. 522W is actually ~ 43A #2 - Total Power as listed is 600W. Most manufacturers will list the total output as the 12V rail capacity. In this case, it should be either marketed as a 550W or a 525W supply - If the label matches the actual components (Other manufacturers will add up the capacity to market the supply as - Either way the label doesn't make any sense) If they can't even manage to calculate the power correctly on the label, who's to say that the power supply is even pumping out what its supposed to be. Beyond that, the label itself is so generic looking that it almost looks like the mfg just tosses the same label onto whatever supply they pick up at a garage sale. The absolute most important component of a computer is the power supply, as everything else runs off of it. It doesn't pay to skimp here. And when you're looking at a powersupply that is (if it is indeed a 585W supply) is 1/2 the price of a 600W supply from a reputable supplier, I'd say there's definitely some skimping going on here... (Even the "style" of the supply looks ancient - This is probably a cheap generic 350W (rated, but probably not even capable of handling that current) supply that's been relabelled as a 600W. And we haven't yet even begun to talk about voltage ripple / drops as the supply approaches its actual limits Here's some sample reviews of generic powersupplies: http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story&reid=71
October 15, 20169 yr Author Interesting....It is a Corsair supply; I thought they were supposed to be one of the best supplies out there? Also - This is the new supply I am looking to buy, but if it seems shotty, I will avoid it.
October 15, 20169 yr Interesting....It is a Corsair supply; I thought they were supposed to be one of the best supplies out there? Also - This is the new supply I am looking to buy, but if it seems shotty, I will avoid it. The link you posted was to a generic power supply But even there, someone at Corsair has no clue how to do the math (And not like its hard.... Watts=Volts*Amps), nor how to add up the totals (looks like they're taking the 3.3Watts and adding it to the incorrect rating on the 12V line to get to 600W) (Although Corsair's website does list the correct values on the 12V line)
October 15, 20169 yr Author Sorry - I see what you are saying now... Yes, the link I posted is to what is currently in the system. I am planning on replacing it with the Corsair Model Shown in the picture, http://www.corsair.com/en-us/cx600-80-plusr-bronze-certified-power-supply The link to the Power Supply Forum thread was trashing on the Corsair CX series, but then there was a list in that same thread that said CORSAIR Builder Series CX500 (CMPSU-500CX) 500W - 34A, supports up to 15 green drives or 9 7200 rpm drives - recommended for medium servers, 10 - 15 drives Is there any particular reason to stay away from the CX Series? I understand the power supply is important, but I'm not in the business of just throwing money at a name. If I can get the performance I need from a Corsair, I don't want to spend an extra $50 on a Seasonic.
October 15, 20169 yr Sorry - I see what you are saying now... Yes, the link I posted is to what is currently in the system. I am planning on replacing it with the Corsair Model Shown in the picture, http://www.corsair.com/en-us/cx600-80-plusr-bronze-certified-power-supply The link to the Power Supply Forum thread was trashing on the Corsair CX series, but then there was a list in that same thread that said CORSAIR Builder Series CX500 (CMPSU-500CX) 500W - 34A, supports up to 15 green drives or 9 7200 rpm drives - recommended for medium servers, 10 - 15 drives Is there any particular reason to stay away from the CX Series? I understand the power supply is important, but I'm not in the business of just throwing money at a name. If I can get the performance I need from a Corsair, I don't want to spend an extra $50 on a Seasonic. Its all anecdotal. I recalled seeing some posts about Corsair CX series and not to use them, and that was the first hit I found on a search.
October 15, 20169 yr Community Expert Sorry - I see what you are saying now... Yes, the link I posted is to what is currently in the system. I am planning on replacing it with the Corsair Model Shown in the picture, http://www.corsair.com/en-us/cx600-80-plusr-bronze-certified-power-supply The link to the Power Supply Forum thread was trashing on the Corsair CX series, but then there was a list in that same thread that said CORSAIR Builder Series CX500 (CMPSU-500CX) 500W - 34A, supports up to 15 green drives or 9 7200 rpm drives - recommended for medium servers, 10 - 15 drives Is there any particular reason to stay away from the CX Series? I understand the power supply is important, but I'm not in the business of just throwing money at a name. If I can get the performance I need from a Corsair, I don't want to spend an extra $50 on a Seasonic. I would personally suggest reading the user's reviews on both Newegg and Amazon for any PS you are considering, focusing mainly on the one-star reviews. Those will tend to give a 'big picture' look at the quality of that unit. I am always leery of anything that has more than 10% one star reviews and would want to carefully read those to see what the true issues are.
October 15, 20169 yr Author Thanks for all the advice everyone. After reading many reviews, and taking in all of the excellent input provided here, I am leaning towards this - https://seasonic.com/product/s12ii-520/ Only concern is the dual 12v Rails. I have read mixed reviews on dual 12v vs single 12v. Is there any reason to worry about that?
October 15, 20169 yr As a general rule single rail supplies are better, but only because it makes it easier to connect everything without having to think about what is on each rail. Generally for dual rails, 12v1 is dedicated for the mobo/cpu 12v2 is everything else including the PCIe power connector if you're using it.
October 15, 20169 yr Author Good to know... But just found on Amazon this - https://www.amazon.com/gp/product/B003HE260I/ref=ox_sc_act_title_1?ie=UTF8&psc=1&smid=ATVPDKIKX0DER Same price, but 620w, modular and single 12v rail.
October 18, 20169 yr Author New Seasonnic PSU came in, plugged it all in, still doesn't facing 2+ minutes to post (pulled out the graphics card and data controller, and down to 1 stick of ram. Tested start up with each disk 1 by 1, all have the same behavior), and still having the crashing behaviors. Its starting to sound like the motherboard to me.... So I guess I have 2 options - scour the internet for a Socket AM2+ motherboard with 4 channels of DDR2 ram, or get a new cpu, motherboard and set of ram. Really annoying, my primary driver to unraid originally over an out of the box NAS option was cheap start up because I had a decently powerful pc lying around. Now I'm going to end up spending more money than I have wanted to on this project.
October 18, 20169 yr You may well be right about the motherboard, but there are 2 more things you can try. - Do a Memtest from the unRAID boot menu, just to make sure your RAM is fine. - Remove the CPU cooler, clean the CPU and cooler surfaces properly, apply a fresh dose of thermal paste, and reinstall the CPU cooler. Anything that crashes under load may be heat related, so it's possible the CPU cooling is not working correctly. Or there's a bridge chipset on the motherboard not being cooled correctly. Make sure all fans are working.
Archived
This topic is now archived and is closed to further replies.