bcbgboy13

http://www.newegg.com/Product/Product.aspx?Item=N82E16822145533&nm_mc=AFC-C8Junction&cm_mmc=AFC-C8Junction-_-na-_-na-_-na&AID=10440897&PID=3332167&SID=u00000687 It is obviously not for Unraid but the performance freaks using RAIDs will be happy. And keep in mind that 4 or 5 of these compressed into a single drive (without extra motors, PCBA etc.) will give us the 4 or 5TB model (and now the Unraid freaks will be happy).

I also have: 1. one used SuperMicro SASLP -SOLD 2. one used BR10i with full size bracket 3. two DELL SAS 6 i/R (one BNIB and one used and flashed with LSI IT firmware) Cables @ $10 each available for all of them. (cables are different)

NORCO RPC-4220 4U Rackmount Server Chassis w/ 20 Hot-Swappable SATA/SAS 6G Drive Bays (Mini SAS Connector) - OEM plus Free Gift: SAMSUNG Spinpoint F4EG HD155UI 1.5TB 5400 RPM 32MB Cache SATA 3.0Gb/s 3.5" Internal Hard Drive -Bare Drive http://www.newegg.com/Product/Product.aspx?Item=N82E16811219033

You clearly have a hardware problems but the forum members are not clairvoyants and you will have to provide the complete hardware configuration as a start. Then there are some inconsistencies in your posts - you are talking about 3 x 2TB WD and then losing 6TB of data - that is clearly not possible in your scenario. And I really do not understand the obsession with that card - any motherboard made in the last few years will have SATA ports that are superior to the ones provided by the SASLP card, so IMHO it will be only natural to start by using the motherboard ports and once you need more to go into additional expenses for the controller and the cables. Since you have only 3 hard drives you can connect them to the MB to eliminate the 5B12 issue with the SM controller and the Forum will try to guide to recover your data. Once you boot - get the HD smart reports, post them here and if you succeed post the syslog.

I believe some of these Zotax mini-ITX with "advertised" 6 SATA ports are kind of misleading as they have only two "real" SATA ports and the others are port multipliers. If this port miltiplier is attached to a single PCIe lane or worst to the PCI bus (so they use old and low cost components) you will suffer in parity checks and rebuilds. It will not show up in a regular RAID system as they will write a stripe to one drive at the full speed, then the next stripe on the next drive again at full speed and so on.. In theory it should not affect Unraid in a regular use (normally read from a single drive) or writing to two drives but make sure the parity is on one of the proper SATA ports. However in parity checks and rebuild situations the performance will be affected as Unraid reads from all disks simultaneously. To somehow mitigate the problem make sure the parity and one of the bigger disks are on the proper SATA ports and connect the smaller ones the the "port miltiplier-ed" SATA ports. Or use an extra controller and do not use these at all.

And the WD Caviar Green 3 TB Internal hard drive - 600 MBps is $105.07 with no tax and free shipping. http://www.electronics-emporium.com/products/Western-Digital-Caviar-Green-3-TB-5400-RPM-SATA-III-64-MB-Cache-Bare%7B47%7DOEM-Desktop-Hard-Drive-%252d-WD30EZRX-%252d%252d-0O004EBEZS601.html Anyone bought something from them?

BR10i will work in an "electrical" PCIe x 4 connector.

You will have to determine your needs first - are you going to stay with 12 HDs (because of case limitations etc.) or not. If you do stay with 12 HDs max then the 650W PSU is simply too powerful and you can safely move to 430W (and even less if available). You have to rate your PSU to provide enough power for all the HDs running simultaneously -power up, data rebuild, parity checks but also if possible to be above the 20% of the PSU rated wattage when your system is idling as under this 20% loading threshold the efficiency usually drops significantly - down to 70 and even 60%... Take a good look on your cooling solutions too- any 120mm fan will consume 2-3W when running (and more if it is power monster). In a typical CM590 case with 12 HDs you will have 3 fans on the front, two on the top, one back, PSU fan and the CPU fan - these fans if running at full speed are easily taking 12-20 Watts. And for example I also have AMD-based DDR2 system as they are great for their build-in ECC support. I have 10 HDs (one of them 3TB) and without any extra modifications (I have all these fans running) and using a 45W TDP 4850e CPU my idling power usage is: - 59 Watts with Corsair 430 - was 74 Watts with PCPQ 610W (I may have had one extra controller inside at that time) Also a property selected motherboard is a must and that usually means do not buy an enthusiastic grade motherboard that will support any CPU and have all the bells and whistles (they still consume power if not used). For the AMD boards look for one that will support only up to 90-100W CPUs - in this way you will get some less power phases that in turn will decrease the power losses in low power (idling state). Added advantage on these boards is that they are also significantly cheaper. And for some PSU recommendations and if money is not the limiting factor take a good look at this: 400 W fanless - for $90 after MIR - http://www.newegg.com/Product/Product.aspx?Item=N82E16817151097 - that will be ideal for your system. 500W fanless - this one apparently is very sophisticated (can shift the power where is needed) and apparently can power up to 20 green drives even been only 500W. - http://www.newegg.com/Product/Product.aspx?Item=N82E16817121083

It has been talked about here - http://lime-technology.com/forum/index.php?topic=15129.0 And there are even lower priced chips like G530 here - http://www.newegg.com/Product/Product.aspx?Item=N82E16819116409

Are you 100% sure that you have ASUS M4A78LT-M According to the Asus specs it uses Realtek® 8112L, 1 x Gigabit LAN Controller(s) And you are not using an Intel card either as according to your syslog you are using Broadcom based LAN: Sep 13 19:58:51 Tower kernel: tg3.c:v3.102 (September 1, 2009) Sep 13 19:58:51 Tower kernel: tg3 0000:03:00.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17 Sep 13 19:58:51 Tower kernel: tg3 0000:03:00.0: setting latency timer to 64 Sep 13 19:58:51 Tower kernel: eth0: Tigon3 [partno(BCM95721A211) rev 4101] (PCI Express) MAC address 00:10:18:18:2c:45 Sep 13 19:58:51 Tower kernel: eth0: attached PHY is 5750 (10/100/1000Base-T Ethernet) (WireSpeed[1]) Bring back this knowledgeable friend and tell him to: 1. Undo the CPU core unlocking if possible and lets hope the mobo and CPU are not damaged by the possible higher TDP. 2. Check out the exact model for the motherboard - should be printed somewhere, also from the white label on the box... 3. Flash the motherboard with the latest BIOS. 4. Load the default values and reboot. 5. Go back into the BIOS and disable any unused hardware features - serial and parallel ports, audio, firewire, even the onbaod IDE controller if you are not going to use any of the older PATA hard drives. 6. Make sure that the onboard SATA drives are set to AHCI mode.( there may be two settings - one for SATA1 to4 and another for SATA5and6) 7. Go into the Supermicro card BIOS and disable INT13H. After all that reboot the computer for an overnight memory test. If successful you will be able to enjoy your Unraid server (but I do not know how good the onboard Broadcom Lan chip is with Unraid)

Your motherboard uses Atheros based LAN chipset. These ones did not play nice with the older Unraid versions, not sure if any development has been done to fix that in 5. Disable any unused hardware feature that you are not using (serial and Par. ports, Audio, firewire, even the VIA-based IDE controller) That will free more resources or prevent them interfering with each other under the new kernel...

According to your signature you are using an AMD Sempron 140 Sargas 2.7GHz and according to the syslog this is transformed into AMD Phenom II X2 545 Processor stepping 02. 1.Have you messed around with the cores? This is one way to end with the instability... And do not forget about the memtest here... 2.The 80GB drive is toasted or does not work well with the controller where is is attached. Try without cache drive.

I have a couple of IBM ServeRAID M1015 LSI 9220-8i PCI-E SAS/SATA RAID 46M0861 controllers for sale together with forward breakout cables SFF8087 to 4 SATA (with latch) Price is $95 shipped Priority mail in continental USA for the controller only. Additional Forward Breakout cables are $10 each (max 2 per card) with no extra charge. International shipping available for an additional $15 (this will cover up to 2 controllers and 4 cables) These are like new controllers as far as I know (factory sealed) and they do come with the full height PCI bracket. SATA 6Gpbs / 3TB+ HDs compatible with flashing to "IT" firmware - more info here - http://lime-technology.com/forum/index.php?topic=12767.0 Full support in the latest Unraid 5 Betas. As there are reported some difficulties flashing these cards you can select to get: - the controller in "as is" condition; - programmed with the latest available IBM firmware; - cross flashed with the latest available LSI "IT" firmware - the Unraid preferred option

What about free local pickup

Over the holiday (here in NA) Intel released 11 new "Sandy Bridge" processors and cut the price of the existing ones: http://www.tomshardware.com/news/intel-sandy-bridge-price-cuts,13355.html Much better choice (from the price point) will be the 35W TDP Celeron G530T ($47) and the 65W TDP Celeron G530 ($42). The price you are going to pay for a retail CPU is going to be higher but these are excellent competition to the AMD Sargas 145. http://www.newegg.com/Product/Product.aspx?Item=N82E16819116409 The only difference is now the ECC memory capabilities for the people that care (and this is not stressed enough in this forum) PS. replaced the wrong link

You have seven HDs with "past failure" attribute set. For Seagate HDs this means that they were run at above 55 deg.C in the past. You probably should take a good look at your cooling especially when doing the "few parity checks" as this is the time when you will have the highest temperatures.

You can thank me later: http://accessories.us.dell.com/sna/products/Power_Surge/productdetail.aspx?c=ca&l=en&s=bsd&cs=cabsdt1&sku=A1833724&~ck=baynoteSearch&baynote_bnrank=7&baynote_irrank=0 The link is for Canada. I believe Dell ships for free... and fast. This is 550VA model - it is perfectly adequate for your system as it is now but it may be limited to 10-12 drives system and slightly less if you connect your modem and router. I you decide on higher power and can wait DELL has some deals from time to time (and I believe with free shipping). And if you want to buy now then check out this - http://www.newegg.ca/Product/Product.aspx?Item=N82E16842101311

Speed test parity check: with Beta-10 (as posted previously): Aug 9 13:07:09 unraid kernel: md: sync done. time=48609sec Aug 9 13:07:09 unraid kernel: md: recovery thread sync completion status: 0 with Beta-12: Aug 29 06:04:10 unraid kernel: md: sync done. time=43557sec Aug 29 06:04:10 unraid kernel: md: recovery thread sync completion status: 0 Both cases - 10 drives system with 3TB Hitachi Green Parity disk.

Hitachi 3TB (as originally announced) were 5 platters design compared to the 4-platters WD

They are plenty of used AMD 2650e CPUs on Ebay for $20-25. This is a 15W TDP part and can run with passive heatsink if you have a larger case (one fan less). However this is only usable on motherboards using DDR2 memories which are now extremely rare. For the AM3 socket they are also very low power CPUs available but they are hard to find (as they were released only as OEM tray devices) and more expensive. There is the 20W TDP AMD 170u (and 160u if you can find them). And if you need a little extra horsepower then you can get the 25W TDP AMD 270u (and the similar 250u, 260u). These will work in an AM2+ socket too. The other beauty of all these is the ECC support (if one choose to care about this and picks a motherboard that will support it) - The Atoms and the low power lower priced Intel offering as of now do not have ECC. However Intel is playing a catch-up and will release new server grade 15W Sandy Bridge chip later this year and sub-10W ATOM chips in 2012. Then you will need to carefully choose the MB (even by examining the picture on the MFR web site) - general rule to avoid anything that will support the 100+W CPUs. Ideally if you can get the 2 phase MB (not sure if there are any for sale) or 2+1 or 3 phase designs. Care must be taken to select the proper case and the cooling - for example the loved Centurion has a 6 120mm fans which when fully loaded will draw 2-3W each. And the biggest problem to meet the lowest kWh is to find a suitable PSU (one must select a model that ideally will run in 20% or more load when idle to be in the optimum power efficiency zone and then be able to handle the start-up current). This means that the cheap Corsair 430 W is not a good choice for a system idling at 40W or less. Unfortunately the choice of good lower powered PSU is extremely limited and they are usually priced higher. Possible alternative are custom PSUs made for DELL (some of them, just some of them are very efficient but it is hard to find which ones and then to find them at reasonable price.)

1.I see that you often have power outages and that during some of them you are losing the network. You should put your router and switch if used on the UPS too. 2. Depending on you configuration (number of drives, PSU and UPS wattage used, state of battery) I suspect you may have some memory errors during these outages. Once booted Unraid is run from the memory and if you get a random bit flipped somewhere in the critical area that may manifest in the way you describe it. There is a reason that every commercial server uses ECC and UPS but the importance of these features is not stressed enough. Unfortunately I am not Linux savvy to say if any of the ECC reporting capabilities are build into the kernel (or can be added in some way) to make it a clear to everyone. And I am getting at least 10 power outages a day myself.

And do not forget they are now available much cheaper dual core Celerons in 1155 - I believe from $70 and from 35W TDP

I did not have any problems running beta 10 (apart from something odd) Regarding the slow speed reported by many: I recently performed "swap-disable" - replaced the old 2TB WD EARS parity disk with precleared 3TB Hitachi Green and then in the same shot used the older parity disk to upgrade 1TB Hitachi data disk. Results: Aug 7 16:45:39 unraid emhttp: writing GPT on disk 0 (sdg) with partition 1 offset 64 Aug 7 16:45:39 unraid emhttp: shcmd (130): sgdisk -Z /dev/sdg &> /dev/null Aug 7 16:45:41 unraid kernel: sdg: sdg1 Aug 7 16:45:41 unraid emhttp: shcmd (131): sgdisk -o -a 64 -n 1:64:0 /dev/sdg |& logger Aug 7 16:45:42 unraid logger: Creating new GPT entries. Aug 7 16:45:42 unraid logger: The operation has completed successfully. Aug 7 16:45:42 unraid emhttp: shcmd (132): udevadm settle Aug 7 16:45:42 unraid kernel: sdg: sdg1 Aug 7 16:45:42 unraid emhttp: copying disk2 to disk0... Aug 7 17:15:41 unraid apcupsd[1445]: Power failure. Aug 7 17:15:43 unraid apcupsd[1445]: Power is back. UPS running on mains. Aug 7 18:31:10 unraid apcupsd[1445]: Power failure. Aug 7 18:31:12 unraid apcupsd[1445]: Power is back. UPS running on mains. Aug 7 23:03:07 unraid apcupsd[1445]: Power failure. Aug 7 23:03:09 unraid apcupsd[1445]: Power is back. UPS running on mains. Aug 8 03:00:08 unraid emhttp: Start array... Aug 8 03:00:08 unraid kernel: mdcmd (45): start SWAP_DSBL Aug 8 03:00:08 unraid kernel: unraid: allocating 54200K for 1280 stripes (10 disks) ... Aug 8 11:08:31 unraid kernel: md: sync done. time=29292sec Aug 8 11:08:31 unraid kernel: md: recovery thread sync completion status: 0 This one resulted in something odd - the interface states that my last parity check was done more than 15000 hours ago (in red color) So I did a parity check: Aug 8 23:36:57 unraid kernel: mdcmd (70): check NOCORRECT Aug 8 23:36:57 unraid kernel: md: recovery thread woken up ... Aug 8 23:36:57 unraid kernel: md: recovery thread checking parity... Aug 8 23:36:57 unraid kernel: md: using 1152k window, over a total of 2930266532 blocks. ... Aug 9 13:07:09 unraid kernel: md: sync done. time=48609sec Aug 9 13:07:09 unraid kernel: md: recovery thread sync completion status: 0 48609 seconds are 810 minutes or app. 61.7 MB/s This is 10 disk system (6 on the MB SATA and 4 on an M1015 card in PCIe x8 port). Unfortunately I do not recall the speed of the previous checks...

I think that is on purpose with many power supply manufacturers and their marketing and sales departments. It is often very difficult to determine how the rails are wired. Some manufacturers do not supply the information at all. They want you to focus on the total wattage number, the bigger the better, even if not useful for your application. But the manual that you linked previously for Enermax 82+ states it supports 11 hard drives for the 385W version and a quad core and a single video card (which the OP does not have) as he is limited to max 10 hard drives anyway with his current hardware (mini ITX MB with two onboard SATA ports and a PCI controller for 8 drives)

Go to www.molex.com Type 79576-3000 in the part number and look at the datasheet (it covers both forward and reverse breakouts) You will need pinout 2 (iPass on CTRL) to SATA Or to get a complete look - www.t10.org/ftp/t10/document.05/05-084r1.pdf