October 3, 20169 yr hey all prior to yesterday i was running a amd x2 250 dual core cpu with 8 gb ram and all was well, no random crashes, could run for a straight month. however last night i made a small upgrade to processing power and threw in a amd phenom 9950 black edition, and usually around 2 hours after boot the system crashes, no access to gui, no access to ftp/ ssh/ telnet, web based apps, etc. and on the display connected to the server is just black, no more console login. a simple reboot brings it back but again it happens after ~2 hours, no heat issues running ~40 deg no load and ~50 load, heatsink is seated and pasted properly i've run memtest 4 times and have no memory errors, im running a 500W power supply with nothing labor intensive like a addon vid card. and tail -f logs does not show anything before crash time. any ideas folks? your help is greatly appreciated
October 3, 20169 yr Author Fix Common Problems plugin in Troubleshooting mode. do i boot into safe mode for that? a little direction would be appreciated
October 3, 20169 yr Community Expert Fix Common Problems plugin in Troubleshooting mode. do i boot into safe mode for that? a little direction would be appreciated Since Fix Common Problems is a plugin, and SAFE mode will not load any plugins, then NO to booting in SAFE mode.
October 3, 20169 yr Author Fix Common Problems plugin in Troubleshooting mode. do i boot into safe mode for that? a little direction would be appreciated Since Fix Common Problems is a plugin, and SAFE mode will not load any plugins, then NO to booting in SAFE mode. yeah i figured out it was a plugin, i never heard of it until now, installed it and ran a scan, deleted some depreciated scripts, and activated troubleshooting mode. will keep you posted.
October 3, 20169 yr Are you sure your motherboard supports the 9950? Two things to check: (a) Does it support 125w CPU's ... you just changed from a CPU with a TDP of 65w to one with a TDP of 125w Did you upgrade your heatsink when you switched CPUs to ensure you have adequate cooling for the higher-power processor? (b) Check and confirm that your BIOS supports the Phenom 9500 => there may have been a BIOS update that added that support, and if you don't have it, that could explain why it's not working properly. In addition, you're using CPU's designed for different sockets !! The Athlon X2 250 uses the AM3 socket; the 950 Black uses the AM2+ socket. I presume that means you're using an AM2+ motherboard, since an AM3 CPU will work in an AM2+ socket, but the reverse isn't true.
October 3, 20169 yr Author Are you sure your motherboard supports the 9950? Two things to check: (a) Does it support 125w CPU's ... you just changed from a CPU with a TDP of 65w to one with a TDP of 125w Did you upgrade your heatsink when you switched CPUs to ensure you have adequate cooling for the higher-power processor? (b) Check and confirm that your BIOS supports the Phenom 9500 => there may have been a BIOS update that added that support, and if you don't have it, that could explain why it's not working properly. In addition, you're using CPU's designed for different sockets !! The Athlon X2 250 uses the AM3 socket; the 950 Black uses the AM2+ socket. I presume that means you're using an AM2+ motherboard, since an AM3 CPU will work in an AM2+ socket, but the reverse isn't true. thanks for your input, my board supports the new processor and is a am2+/am3 motherboard, the date code on the 9950 is actually older than the x2 250 edit: i forgot to add, the heatsnk im using keeps the phenom at around 50 C under laod, my research shows the cpu is capable of hitting 60, so i think im ok on the hs/fan.
October 3, 20169 yr The date code isn't really what's important -- BIOS support for the CPU is what matters. Although I agree that if it's an older processor, then if the BIOS supports the X2 250 it almost certainly is a new enough BIOS version to include 950 support. Certainly wouldn't hurt to check the CPU compatibility list for your motherboard, however. You said you ran MemTest 4 times (since yesterday) ... clearly you're not letting it run for a lot of cycles. Wouldn't hurt to run it overnight just to be sure this isn't a memory related issue [Run it for at least 8 hours]. Is your power supply a high-quality unit? [80+ certified, Active PFC, etc.] I've seen a lot of "quirky" issues with low-end inexpensive power supplies that will instantly resolve with a better unit. Lower end units simply don't have outpus as "clean" as the higher quality suppies do.
October 3, 20169 yr Author The date code isn't really what's important -- BIOS support for the CPU is what matters. Although I agree that if it's an older processor, then if the BIOS supports the X2 250 it almost certainly is a new enough BIOS version to include 950 support. Certainly wouldn't hurt to check the CPU compatibility list for your motherboard, however. You said you ran MemTest 4 times (since yesterday) ... clearly you're not letting it run for a lot of cycles. Wouldn't hurt to run it overnight just to be sure this isn't a memory related issue [Run it for at least 8 hours]. Is your power supply a high-quality unit? [80+ certified, Active PFC, etc.] I've seen a lot of "quirky" issues with low-end inexpensive power supplies that will instantly resolve with a better unit. Lower end units simply don't have outpus as "clean" as the higher quality suppies do. yeah i hear you on the date code, but yes, i have checked the official datasheets, i will link you the board im using and everything checks out hardware compatibility wise, the power supply is not a cheapo-came-with-case psu, however im eating supper and not about to run down 2 flights of stairs to check the brand name, and yes, its a known brand, just cant remember which. http://www.gigabyte.com/products/product-page.aspx?pid=3193#ov
October 3, 20169 yr Both CPU's are indeed on the CPU support list => the 9950 is supported from stepping B3, but I doubt that's an issue (stepping should be stamped on the CPU). Name brand isn't the only thing to check r.e. the PSU => many good manufacturers have low-end units they build to satisfy the low-cost market, and these can reflect compromises that can cause issues with some components. e.g. Corsair makes superb power supplies in their HX and AX lines; and some very good mid-range units; but also sell an inexpensive CX series that can cause issues with many systems. First thing I'd do, however, is an overnight run of MemTest-86 with all 4 cores active (You can set this in the options => not sure if UnRAID is using the latest version, so you might want to boot from a current download: http://www.memtest86.com/ ).
October 4, 20169 yr Author attached is the diags and syslog generated by the fix common problems plugin in troubleshooting mode. crash leads to everything unresponsive, and console login screen on server is blank streamflix-diagnostics-20161003-2058.zip
October 4, 20169 yr Author Both CPU's are indeed on the CPU support list => the 9950 is supported from stepping B3, but I doubt that's an issue (stepping should be stamped on the CPU). Name brand isn't the only thing to check r.e. the PSU => many good manufacturers have low-end units they build to satisfy the low-cost market, and these can reflect compromises that can cause issues with some components. e.g. Corsair makes superb power supplies in their HX and AX lines; and some very good mid-range units; but also sell an inexpensive CX series that can cause issues with many systems. First thing I'd do, however, is an overnight run of MemTest-86 with all 4 cores active (You can set this in the options => not sure if UnRAID is using the latest version, so you might want to boot from a current download: http://www.memtest86.com/ ). heres the PSU http://www.thermaltake.com/products-model.aspx?id=c_00001655 will run memtest next after troubleshooting mode diags get reviewed
October 4, 20169 yr This might be wrong, but did you use the PSU calculator recently? Since you switched out the CPU, the power budget you need may have expanded. In that case you will need a new PSU. I'm no expert, but it sounds like the issue I had with my workstation where it stopped working after a certain period. Turns out that the PSU's performance had decreased to about 80%, and when I hit the max load it had just broken loose. You mentioned having no access to any console or webUI, so I'm betting very heavily towards the PSU power budget. Especially since your RAM sticks with the previous CPU configuration ran fine.
October 4, 20169 yr Author This might be wrong, but did you use the PSU calculator recently? Since you switched out the CPU, the power budget you need may have expanded. In that case you will need a new PSU. I'm no expert, but it sounds like the issue I had with my workstation where it stopped working after a certain period. Turns out that the PSU's performance had decreased to about 80%, and when I hit the max load it had just broken loose. You mentioned having no access to any console or webUI, so I'm betting very heavily towards the PSU power budget. Especially since your RAM sticks with the previous CPU configuration ran fine. http://outervision.com/b/TY9ePg
October 4, 20169 yr Author just adding to the thread that im currently updating to the unraid release released today. awaiting review of diags just some thoughts, could there be bios settings that had no issues with the x2 that the x4 does not like? i reset to optimized defaults but that didn't seem to matter.
October 4, 20169 yr I'd also lean towards the PSU => that's a very low-end unit and may simply have issues with the higher power demands of the new CPU, or the stability of the power buses might be more criital with the higher end processor.
October 4, 20169 yr Author I'd also lean towards the PSU => that's a very low-end unit and may simply have issues with the higher power demands of the new CPU, or the stability of the power buses might be more criital with the higher end processor. it used to run a beefy video card back in my pc gaming days without issues. im doubting the psu is the issue as i ran the calc and even at low end running at 90% usage on the phenom id only need a 400 watt one, this one can peak 600w and has 80 cert, and APF , wasnt low end when bought lol. dont get me wrong im not trying to undermine you, i will take all the help/advice i can get right now, i just dont suspect the ram or PSU is the issue. if anything i wouls suspect the issue is with the CPU, as i did buy it from ebay. my setup ran flawlessly before upgrading it. im aware of the wattage difference between cpu's but 500watts for a micro atx board and 4 harddrives is totally underpulling on the psu imho
October 4, 20169 yr Did you run MemTest for an extended time? An 8-10 hour run of MemTest would not only thoroughly test the RAM, but would also confirm that the CPU is running okay -- and provide some confidence that the issue may in fact be a configuration issue. How many memory modules do you have installed?
October 4, 20169 yr Author Did you run MemTest for an extended time? An 8-10 hour run of MemTest would not only thoroughly test the RAM, but would also confirm that the CPU is running okay -- and provide some confidence that the issue may in fact be a configuration issue. How many memory modules do you have installed? i just edited my post above, as for running memtext extended, i was waiting for my server to crash so i can post the logs and have them reviewed first, if theres nothing in the logs, next step was a half day run on memtext with 4 cores active, and i have dual 4gb ddr2 sticks kingston branded
October 4, 20169 yr Author another side note.. im just typing as i think.. but since i lose video on the console display, could the issue be video driver related? i ask because reading the forums, usually when unraid crashes theres still video on the console, wether it be kernal panic error codes or a lock up of text on the screen. now i am running from the integrated graphics, im just putting things out there for now
October 4, 20169 yr Pull two memory modules and see if it's more reliable with only 2 modules. The bus loading with 4 unbuffered memory modules is very high => may have not been an issue with the lower end CPU, but may be causing issues with the higher performance processor. If that does the trick, you may be able to reliable use 4 by adjusting the memory timings (adding an extra latency cycle and/or slightly increasing the voltage); but any unbuffered system will be more reliable with lower loading. I NEVER install 4 modules in systems that use unbuffered modules.
October 4, 20169 yr Author Pull two memory modules and see if it's more reliable with only 2 modules. The bus loading with 4 unbuffered memory modules is very high => may have not been an issue with the lower end CPU, but may be causing issues with the higher performance processor. If that does the trick, you may be able to reliable use 4 by adjusting the memory timings (adding an extra latency cycle and/or slightly increasing the voltage); but any unbuffered system will be more reliable with lower loading. I NEVER install 4 modules in systems that use unbuffered modules. i think you may have interpreted the post incorrectly, there are two slots, i have dual sticks of 4GB memory for a total of 8GB
October 4, 20169 yr I just recently had a similar issue when I upgraded my system with new RAM and a new CPU. After the upgrade I would get random errors and freezes.. The CPU was approved for my motherboard. Latest BIOS. I started playing around with Memory and ran tests for hours.. nothing worked. In the end I tried a different CPU and never had a single issue after.. I still don't know why that CPU wouldn't work properly in there..
October 4, 20169 yr Author I just recently had a similar issue when I upgraded my system with new RAM and a new CPU. After the upgrade I would get random errors and freezes.. The CPU was approved for my motherboard. Latest BIOS. I started playing around with Memory and ran tests for hours.. nothing worked. In the end I tried a different CPU and never had a single issue after.. I still don't know why that CPU wouldn't work properly in there.. how about your console login screen, would it black out or just freeze text?
October 4, 20169 yr I just recently had a similar issue when I upgraded my system with new RAM and a new CPU. After the upgrade I would get random errors and freezes.. The CPU was approved for my motherboard. Latest BIOS. I started playing around with Memory and ran tests for hours.. nothing worked. In the end I tried a different CPU and never had a single issue after.. I still don't know why that CPU wouldn't work properly in there.. how about your console login screen, would it black out or just freeze text? I believe it would just freeze text.. but I am using the onboard video I still had all the old parts that have worked fine for years, so I just started swapping stuff around till it could only be the cpu.. New RAM + CPU = fail, Old RAM + CPU = fail, new RAM + old CPU = ok. Ordered a different CPU and it worked fine as well.. I'm not saying you have this issue, just something that happened to me..
Archived
This topic is now archived and is closed to further replies.