One2go

Directory Opus $88 US you got to be kidding. Not for me, you don't happen to have any financial vested interest in that utility do you?

Thanks for the advice. After carefully reading all 18 pages of the 5.Beta 12a Announcement I am taking my time updating the server. I only noticed the frustration with the BLK errors of the 8 port Supermicro controller. I can live with the spin up of the discs and I only turn my server on when needed I don't run it 24/7. Thanks again for your help.

Thank you your comments are MOST helpful. I will only have ONE of the SASLP-MV8 Controllers & I disabled the Atheros on-board LAN replaced it with an Intel NIC which runs smoothly. I just ran a parity test and all looks excellent with no errors or other abnormalities. Excellent suggestion regarding the additional parity check before swapping the parity drive. Also did not consider a backup of the flash drive. Is this a simple matter of copying the entire drive OR is there a backup utility or procedure for this? Keep in mind my Linux skills and knowledge are very limited as I often have to search for examples of Commands from the prompt. I seem to recall that there is a problem with the PreClear script for 3TB drives or has this been solved or is there a working custom PreClear script for 3 TB drives? Thanks for your help and I'll surely read the Beta threads diligently.

Thanks did not know that about the Atheros and 4.7 which I am running. If I did, I would have installed a new NIC longtime ago. As it turns out with the new Intel NIC my setup is now very much to my liking.

Your reasoning about HDParm has crossed my mind and it is one of the reasons that I am leaning towards the updating to the 5 beta version. Still would like to know if there are any other abnormalities with the beta version. This is a job that you just don't jump head over heal into 16TB of collected data are begging to be handled gingerly.

The drivers for the NIC are in the Linux distribution, so I guess you will not find out until you plugged it in. Intel drivers are always included as their NICs are the defacto standard. If you have 1 free 1 lane PCI-e slot left then just look at this glowing report of this card http://www.newegg.com/Product/Product.aspx?Item=33-106-033 (230 reviews and still 5 egg ratings). Cost you $28 with free shipping and mine arrived 3 days after ordering. Or if you have a PCI empty slot check this card out http://www.newegg.com/Product/Product.aspx?Item=N82E16833106121 (430 Reviews & still 5 Egg rating) This was also recommended by someone else here at the forum $31 with free shipping.

I have followed bjp999, ProStuff1 & your comments closely on the 2TB to 3TB move and are still tossing around the criteria of available time versus risk of running a beta release on the server that is in use daily. The beta version that supports 3TB drives has been around for a while now, besides the spin-up glitch are there any other show stoppers you are aware off, as I have read that some feel the latest beta release is quite stable?

The speeds I mentioned is writing straight to the array, not a cache drive and not reading and those speeds are pretty much in line with the performance under UnRaid 4.7 when calculating parity as it is written to the array.

Check the PCI NIC recommendation in this thread http://lime-technology.com/forum/index.php?topic=15753.0 Do yourself a favor and stick with Intel it will ensure that the Linux distros will have drivers for your NIC. You will not regret it, for with Intel NICs you just install and forget about it, they just work. With the aIntel NICs you should be spending between $27 & $30. Switches are very cheap, TrendNet, DLink all in the order of $30 to $80. With Jumbo frames it is not a panacea, but all the pieces in the network should support it with the most important pieces being the end parts, server & front end (HTPC). My Win 7 machine's NIC an Intel has only 3 MTU settings 4088, 9014 & disabled. Shoot for the 9014 and it will suffice.

The Mother Board is ASUS M4A78-E motherboard It has as the NIC the Atheros L1E Gigabit LAN controller featuring AI NET 2 here is what one guy wrote in the NewEgg feedback:

I am standing in front of doors 1, 2 & 3 scratching my head how to proceed with the disc upgrades on my production server and have a few questions, even after spending hours researching and reading forum posts of the experience others had with 3TB drives. Server Hardware: Mother Board ASUS M4A78-E motherboard Adaptec 1430SA Controller Supermicro AOC-SASLP-MV8 Controller AMD 790GX / SB750 Chipset for 5 SATA ports 9 Samsung 2TB drives used in the array 4 Hitachi 3TB drives on the bench waiting to be used. The 3TB drives will be connected to the Supermicro controller with the exception of the parity drive which I would like to connect to the Mobo. Is there an advantage to this? Door #1 Should I upgrade from UnRaid version 4.7 to version 5 Beta 12a on my server so as to be able to make use of the full 3TB size? If so is there a modified PreClear script and does it work with 3TB drives? This would be the preferred option but it depends on how stable the release is. Door #2 Use the HDParm procedure to reduce the drives to 2TB and then integrate them into the array and use version 4.7. I can see where this maybe the saver method but also in the end more time consuming. Are there any hiccups or precautions to be considered when expending from 2TB to 3TB using the HDParm command? Door #3 Be patient wait until a stable version 5 release and use 2TB drives one at a time to cover the needed storage space? This requires the least amount of changes but is also the most expensive as the 2TB drives will be spare parts eventually. Possible usage in a test server. Not my favorite Door. Any experience that you would like to share is much appreciated. It might be a good idea to start a sticky with a collection of steps to upgrade from a box with 2TB drives to a box with 3TB drives. Thanks O2G

Well I had this up for s few days & in the first post is the request for a NIC recommendation, so after a few days when none were forthcoming I ordered the PCI-e card which works like a charm & I couldn't be happier. For me no need for additional ports as the Mobo has One 8xPCI-e & One 16xPCI-e slot plus Two 1xPCI-e slots. Have ports for 17 drives but no more space for drives unless I change the 4 in 3 to 5 in 3 adapters. But thanks for the recommendations as someone else may benefit from it.

When I rebuild my parity it said something in the order of 1600 went to bed 8 hours later it said 18 minutes left. Initial calculation I guess doesn't include speed increase down the road. It is at best a rough estimate.

Well here we go, continuing with the monologue. I installed the Intel EXPI9301CTBLK Network Adapter 1000Mbps and disabled the on-board LAN. Well what a difference and I now understand the glowing reports & comments on NewEgg about this NIC. Set MTU=9014 with no problem. No more hesitation and waiting when copying and moving data to the array. TeraCopy reports that while moving directly to the array a 1.5 Gig file speeds between 42 MB/s and 47 MB/s. Nice one, glad that this part is over. Now to the bigger job the upgrade path to 3TB drives but that I will post in another thread. O2G

No this is definitely not correct nor normal. I noticed from your log that you are having a Realtek chip set for the NIC. eth0: Identified chip type is 'RTL8168C/8111C'. Search for that on the board here because that chip set is nothing but problems and it is documented. Do yourself a favor and buy for $30 a decent Network Interface card. Look at the second post for a recommendation of a card http://lime-technology.com/forum/index.php?topic=15753.msg146714#msg146714 I just installed that card and it is a difference between night and day. Also be mindful that if you do the operations from a Windows 7 machine using Windows Explorer it sucks. Just Google for "Slow Copy & Move" and you will see the many different people that complain about how Windows is really slow when interfacing with Linux for Copying or Moving. Use TeraCopy, RSync or some other program for copying. If you go to the Console you should have blazing speed. Another way to test or narrow down the problem, use Putty or some other program that gives you a session and use the native Linux commands and you will know if it is a problem with the machine. My speed increase writing to the array after switching the NIC is remarkable.

I am facing the same decision. There are plenty of threads here discussing this approach. I will wait until a stable 5.* version is released that supports 3TB drives. I have 4 of the Hitachi drives and am considering to do the HDParm trick of making a partition equal to 2 TBs. Then when the stable 5.* version is released increase the partition to 3TB and let the array do its rebuilding. I think this is going to be a lengthy process time wise.

Well I thought I eliminate for $30 the possibility of the on-board NIC being dodgy. I got this Intel one from NewEgg http://www.newegg.com/Product/Product.aspx?Item=33-106-033. Users stated that it works very well with Linux and one user made this interesting observation regarding Jumbo Frames: I guess I spent $30 to answer a few of my own questions and will find out if my setup will allow for a MTU=9014 setting. Which still leaves one unanswered question, in the version 5.0 beta releases have there sofar been unpleasant experiences with this type of NIC & with Jumbo Frames in general? Thanks O2G

First off my setup: This is my server configuration: http://lime-technology.com/forum/index.php?topic=5102.0 On the network I am using a managed switch 3Com 3CDSG8 with Jumbo Frames enabled and 1000 Mbps speed at the server side in the server room. QOS is disabled. I am using a D-Link DGS-2208 8-Port 10/100/1000 Desktop Switch that supports Jumbo frames at the HTPC side in another room. Both switches are connected via cat 6 app. 80 ft. HTPC is a Win 7 32 bit machine & the workstation is a Win 7 64 bit machine both set to use Jumbo Frames MTU=9014. In addition I have a QNAP NAS on the network with Jumbo Frames enabled. This leaves the Unraid server on normal frames with the ifconfig eth0 command indicating MTU=1500. In addition connected to the HTPC side switch is the HTPC, a Denon receiver & a Panasonic PJ and connected to the server side switch besides the cable connecting the two switches is the QNAP NAS, Unraid Server, a Workstation plus two 100 Mbps devices (Comcast SMB cable modem and a Sonos bridge). I searched the forum here and got different opinions on the topic of Jumbo Frames. Speed increase in one case and no difference in a another case because on a Giga Byte LAN most of the time is spent waiting for the next packet. Wanting to make sure that I get the max read speed I thought I experiment with Jumbo Frames. At present my Write Speed to the Array is between 20 - 39 MB/s depending if it is from the workstation or from the NAS. I am using TeraCopy to get an idea of the transfer speed. I also understand that my read speed it much higher, I have not measured / read the read speed yet. The main reason why I started to look into using Jumbo Frames was the length of time it took before TeraCopy started the actual copying to the array and in a few instances Win Explorer complained that the network location was no longer available. I realize that there maybe another cause for it. So from the command prompt in Putty I entered the ifconfig eth0 mtu 9000 command. It complained and I found out that the on-board network will only go as high as mtu=8170. So I used different mtu sizes 8170, 8014, 8000, 7000 & 6000 but in each instance it actually worsened the network availability, to the point that Putty would no longer allow me to connect but even worse the Web Interface to the main page that allows me to reboot and shutdown the server just got stuck on the message transferring data from Tower. UnMenu did load but when trying to switch to UnRaid main from UnMenu it again got stuck. I shutdown the server from the console and with UnRaid mtu=1500 all worked again. So none of the Jumbo Frame settings worked for me on the UnRaid server and I am back to mtu=1500 for UnRaid and all other devices are set to Jumbo Frames. Here are a few questions and request for help. 1. Since I have two 1 lane PCI-e slots available would it be worthwhile to disable the on-board network connection and use a NIC that can set Jumbo Frame size to mtu=9014? 2. If so, are there any recommendations for a NIC? 3. If I use a separate NIC does UnRaid provide the drivers & thus it has to be a tested and approved NIC? 4. Is the driver for the on-board network part of the UnRaid OS or part of the BIOS? 5. Since this is an Asus Mother Board is there a need or way for updating the MB to get Jumbo Frames with mtu size 9014 from the on-board NIC? 6. Or is the whole idea of using jumbo frames not worth the hassle because of the speed increase being negligible? 7. Do the 5.0 beta releases make any changes to the usage of Jumbo Frames? (I could not find anything in the Change logs.) Any help would be very much appreciated as I am trying to get this server fine tuned because this coming week I will try to prepare the box that has 8 empty slots for migration to 3 TB drives plus installation of the Supermicro AOC-SASLP-MV8 controller. Thanks O2G

I did not use the JP firmware. Afteer reading through this thread I endorse some of the opinions expressed such as: It just doesn't sound correct to issue two different flash programs for different drives distributed in different localities and then Samsung is telling me it doesn't matter. Someone is having their thumbs where should not be. But I guess that this is history since the Samsung HDD division is no more and has been sold. I take my chances with using the drive as is and keep a closer look on it. When 3 TB drives are supported the drive is history anyhow.

Bought one of these Samsung HD204UI using the Newegg deal manufactured in 2011. On the disc label I found 2011/01 and I interpreted that as the manufacturing date. Nowhere else did I see anything that looked like a manufacturing date nor did unMenu or the Unraid main menu tell me anything about date of manufacturing. This is part of what unMenu reported: SMART status Info for /dev/sdi smartctl 5.39.1 2010-01-28 r3054 [i486-slackware-linux-gnu] (local build) Copyright (C) 2002-10 by Bruce Allen, http://smartmontools.sourceforge.net === START OF INFORMATION SECTION === Device Model: SAMSUNG HD204UI Serial Number: S2H7J1CB121131 Firmware Version: 1AQ10001 User Capacity: 2,000,398,934,016 bytes Device is: Not in smartctl database [for details use: -P showall] ATA Version is: 8 ATA Standard is: ATA-8-ACS revision 6 Local Time is: Wed May 4 20:56:48 2011 EDT SMART support is: Available - device has SMART capability. SMART support is: Enabled Installed it, unplugged the power from my other 8 drives of which 4 were connected to the mobo & 4 were connected to the Adaptec 1430sa controller. Created a bootable USB key using Hiren's Boot CD and copied the Firm Ware flash to the USB drive. Booted the machine and ran the FW flash. It found the 4 HD203WI drives connected to the Adaptec controller and complained that this patch is not for those drives. It then tried to apply the patch to the HD204UI but came back very quickly with no warning messages or Firm Ware was successfully flashed or any error message. Ran the Firm Ware flash again, same results quick exit from the flash program. I interpreted that as Firm Ware was patched and up to date. I didn't want to go through the hassle and actually test the drive with the condition that causes the silent corruption and thus proceeded with the preclear, hope this will not come back and bite me in the you know what. It took almost 26 hours and this is what preClear reported: ================================================================== 1.9 = unRAID server Pre-Clear disk /dev/sdi = cycle 1 of 1, partition start on sector 63 = Disk Pre-Clear-Read completed DONE = Step 1 of 10 - Copying zeros to first 2048k bytes DONE = Step 2 of 10 - Copying zeros to remainder of disk to clear it DONE = Step 3 of 10 - Disk is now cleared from MBR onward. DONE = Step 4 of 10 - Clearing MBR bytes for partition 2,3 & 4 DONE = Step 5 of 10 - Clearing MBR code area DONE = Step 6 of 10 - Setting MBR signature bytes DONE = Step 7 of 10 - Setting partition 1 to precleared state DONE = Step 8 of 10 - Notifying kernel we changed the partitioning DONE = Step 9 of 10 - Creating the /dev/disk/by* entries DONE = Step 10 of 10 - Verifying if the MBR is cleared. DONE = Disk Post-Clear-Read completed DONE Disk Temperature: 28C, Elapsed Time: 25:45:21 ========================================================================1.9 == SAMSUNG HD204UI S2H7J1CB121131 == Disk /dev/sdi has been successfully precleared == with a starting sector of 63 ============================================================================ ** Changed attributes in files: /tmp/smart_start_sdi /tmp/smart_finish_sdi ATTRIBUTE NEW_VAL OLD_VAL FAILURE_THRESHOLD STATUS RAW_VALUE Raw_Read_Error_Rate = 100 252 51 ok 0 Hardware_ECC_Recovered = 100 252 0 ok 0 Multi_Zone_Error_Rate = 100 252 0 ok 0 No SMART attributes are FAILING_NOW 0 sectors were pending re-allocation before the start of the preclear. 0 sectors were pending re-allocation after pre-read in cycle 1 of 1. 0 sectors were pending re-allocation after zero of disk in cycle 1 of 1. 0 sectors are pending re-allocation at the end of the preclear, the number of sectors pending re-allocation did not change. 0 sectors had been re-allocated before the start of the preclear. 0 sectors are re-allocated at the end of the preclear, the number of sectors re-allocated did not change. root@Tower:/boot/Scripts# Followed the instructions on page 1 for preClear to add the drive to the array. So far so good and all seems well in the land of UnRaid with a Samsung F4 drive. One final word, a big thanks to Joe & Tom for making working with this product a delight, especially for those who are Linux noobs like me.

I have 8 of these drives in my server. 4 run of an Adaptec 1430SA card and 4 run of the mobo. Bought one additional drive as it fits in the case and mobo has 5 ports. I plan eventually to add 8 more into the Antec 1200 case. Whenever I want to watch something I turn the beast on and when done I turn it off. I have had it on for several days but did not notice the temperature problem. The ambient temperature in the room can go up to +80F but again so far runs like clock work.

Got one and received a notice that this drive is being considered for award of the month, please vote.

Ditto here, I have about 9 Samsung 2TB drives and a slew of 500 Gig to 1TB drives for over 5 years and have yet to return a drive. The worst that happened was the partition changed from static to dynamic after a mobo BIOS update. R-Studio to the rescue and I got the data off. Reformatted the drive and it is still operational.

I just bought 4 of them for $146.99 using BING to buy them and just finished preclearing them. Love those drives, just hope that they will last.

The PreClear of the 4 new drives finished and the array is back on line. Just something funny with the PreClear final summary. Here is what showed on the Putty screen: ============================================================================ == == Disk /dev/sda has been successfully precleared == ============================================================================ S.M.A.R.T. error count differences detected after pre-clear note, some 'raw' values may change, but not be an indication of a problem 55c55 < 1 Raw_Read_Error_Rate 0x002f 252 252 051 Pre-fail Always - 0 --- > 1 Raw_Read_Error_Rate 0x002f 100 100 051 Pre-fail Always - 57 67c67 < 195 Hardware_ECC_Recovered 0x003a 252 252 000 Old_age Always --- > 195 Hardware_ECC_Recovered 0x003a 100 100 000 Old_age Always 72c72 < 200 Multi_Zone_Error_Rate 0x002a 252 252 000 Old_age Always - 0 --- > 200 Multi_Zone_Error_Rate 0x002a 100 100 000 Old_age Always - 2 ============================================================================ This summary was identical on all drives with the exception of the very last column. Any idea why that maybe or what this means?