April 10, 201115 yr I got the items before Christmas and built this exact setup, just got it all together and unRAID installed and working http://lime-technology.com/forum/index.php?topic=9644.msg92334#msg92334 I have 2 2tb WD ears drives in there now. I have parity OFF at the moment though, I was going to start copying the data first then turn on parity afterward to let that run over night (heard it was faster to put the initial data on there and then sync up). I believe the NIC in that mobo is 1gig, not sure why it wouldn't be. Even at 100mb I would be getting something better than the 300k I am getting (yes, 300k). Initially I had the Tower across the house in an empty room (noise) and I thought it might have been my CAT5e run (terminated bad, etc). So now it is sitting next to the source computer, both connected to a 1gig switch. Still getting the slow copy speed (300k) and after about 10 minutes it says the destination is unavailable and disconnects. The https:\\tower interface seems *pretty* responsive. It doesn't lock up or anything. Booting the unRAID server seems pretty uneventful. It takes a few minutes and its available on the web interface. Going through the tabs/pages on the interface I don't see any errors. Though maybe I don't have it setup completely correct. New to unRAID all together though I am in the IT world by trade, Windows, Cisco, etc - no *nix experience really. I will try to list what I hope is relevant: - you can see my hardware config with the link posted at top. - drives are new, no jumpers (per the sticky at the top of the forum). The parity disk (disk0 I guess) is off, but in there - no sync. The other drive is MBR: 4k-aligned file system reiserfs. I didn't do anything command-line to prep the drives or anything. I just stuck them in out of the OEM box and used the web-interface to make the changes per what I had read on the forum. - the shares are Movies and TV Shows. MOST of the boxes under 'User shares' are empty... I typed in '2' for split level and 17000000 for Min.free space (largest files are around 8gig, so I doubled that +1g per some things I read in the user manual online). Both are 'High-Water' and Export SMB is export read/write. Export NFS is blank as well as Inc and Exc disks (no drop down menus, just a blank text box???) - 'Export Settings' on the top of the shares page is Export read/write for Flash/SMB Disk/SMB, shares NFS is BLANK and users shares is 'enabled' - On the Setting tab it has 'simple' for the 'Share security' Just realized now that I am getting prompted when I am on my Mac and Windows machines when connected, which I would assume would be correct behavour. However, it asks me each time, several times and I just realized I am putting in some standard credentials I use that ARE NOT on the unRAID server... hmmm. I just have the root account setup (as far as I know) on the unRAID server. Is there anyone that has any idea? Something that I am missing? I can provide more information if its need of course (I may need help with the commands if I need to do something from the terminal) Thanks!
April 10, 201115 yr Author Wanted to add that I am interacting with the shares and have created some folders, etc. Just a file of any size copied to just doesnt cut it. UNC\SMB navigation on the WIn 7 and Mac machines is slow to practically non-existent. \\tower\ EVENTUALLY comes up, but often just times out. Basic network connectivity is there, the web interface comes up and there isn't an IP conflict or anything like that. Version is 4,7 Pro and here is the syslog syslog.doc
April 10, 201115 yr Sounds like some sort of network/nic issue. Can you try disabling the onboard and add in a PCI nic to test? Try different ports on the switch? Shawn
April 10, 201115 yr It'd be good to see the syslog and how your nic is negotiating its connection. Maybe try a different switch? Or just a cable to the client system?
April 10, 201115 yr Author Thanks for looking, guys. Appreciate it. I don't have a NIC on hand, I probably have several in the office, probably just 10/100's though (certain of that, actually). If I picked up a 10/1000 PCI NIC is there any brand that works with unRAID best? The router/switch is an Apple Airport Extreme and I am using DHCP on the unRAID, I might try setting it to static maybe. Of all the home office type AP/Routers/Switches the Apple Airport has been the most solid - never have to mess with it. I am hoping it's not that. I have a gig switch that is downstream, I can tap the Tower into that, see if that helps. Anything in the log about the NIC interaction that stands out? s
April 10, 201115 yr From the console or Telnet, run this: root@Tower:~# ethtool eth0 Settings for eth0: Supported ports: [ MII ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Advertised auto-negotiation: Yes Speed: 1000Mb/s Duplex: Full Port: MII PHYAD: 1 Transceiver: external Auto-negotiation: on Supports Wake-on: g Wake-on: d Link detected: yes root@Tower:~# And should give you back the relevant info to your nic. Want to see what the "speed" and "duplex" settings are. I wonder if it is auto detecting some weird setting, like 10/Half. For picking up a Nic, you could check the Wiki for supported hardware that has been tested...
April 10, 201115 yr Try reading the sticky thread at the top of this forum that's titled include your syslog and version. Peter
April 10, 201115 yr Author Try reading the sticky thread at the top of this forum that's titled include your syslog and version. Peter You sound flippant. That's fine. Try reading the 2nd post. Had issues attaching the file so I replied to myself and added more information. Apologize if that isn't protocol.
April 10, 201115 yr Author Thanks for the telnet commands. Not home now but I will give it a shot when I do get home. Leaning towards nic myself. The web UI seems to be ok. But any unc interaction locks up my machines (never reboot my Mac and navigating the shares would completely lock up my machine). I remember reading that you can transfer data via a drive connected to the unraid box. I might try that and see if the data copies quickly. Would be really surprised if it's the controllers or HDD's. The mobo # I will post. I can't recal at the moment. I did choose one because a similar one had NiC issues - maybe I got confused and choose the wrong one. My base build was from Raj's list - but thing had went out of stock so we substituted. Really really appreciate the good help. Sean
April 10, 201115 yr What are you copying from/to? From a Windows box to a drive mapped to the unRaid? To a UNC path? I know the copying from within Windows using Explorer is about the slowest, all the overhead. I use a program, called Beyond Compare, that has many options and makes moving data to my unRaid from windows a LOT faster. From within unRaid you can also use Midnight Commander to move data between disks. Again, on console just type "mc". If you use putty form a Windows client, you might even get mouse support with MC. Shawn
April 10, 201115 yr Your parity drive can't be "off" because it appears a parity sync starts and then is canceled each time the array starts.
April 11, 201115 yr Author What are you copying from/to? From a Windows box to a drive mapped to the unRaid? To a UNC path? I know the copying from within Windows using Explorer is about the slowest, all the overhead. I use a program, called Beyond Compare, that has many options and makes moving data to my unRaid from windows a LOT faster. From within unRaid you can also use Midnight Commander to move data between disks. Again, on console just type "mc". If you use putty form a Windows client, you might even get mouse support with MC. Shawn Copying from UNC path on Windows and on a Mac. The Mac locks up - the Win 7 laptop comes back around after a few minutes. Thought I had a 10/100 NIC in an old tower in the garage, it's onboard though - so I will have to wait until Monday to grab one from work. The mobo is an Asus M4A78LT-M, which is the 'correct' one according to this post: http://lime-technology.com/forum/index.php?topic=6304.0;wap2 Going to try some of the telnet commands now: Settings for eth0: Supported ports: [ TP MII ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Advertised auto-negotiation: Yes Speed: 1000Mb/s Duplex: Full Port: MII PHYAD: 0 Transceiver: internal Auto-negotiation: on Supports Wake-on: pumbg Wake-on: g Current message level: 0x00000033 (51) Link detected: yes Just loaded TeraCopy on the Win7 laptop and started moving a 7gig ISO and the rate was 410kb/s Short of trying another NIC tomorrow I am not sure what more I can change.
April 11, 201115 yr Now that is weird... Check mine again, specifically the link modes: root@Tower:~# ethtool eth0 Settings for eth0: Supported ports: [ MII ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full The GB.... mine only advertises and supports Gb in Full... no half, that is because GB, far as I have read, only supports full duplex on ethernet. A lot of switches also only support GB running at full duplex. Cisco was one of the first to not include any half duplex support for GB. Weird that your card says it supports and advertises GB full and hald duplex modes. Shawn
April 11, 201115 yr Author I wasn't aware of the full/half issues on 1Gb cards. At work, all the servers and clients all operate at full at 1gb - when they don't I guess I know we have a problem Well, I spent a good part of the day loading unMenu, the SMTP relay, Screen, a network performance monitor, and a couple other packages that looked interesting. Everything seems to be installed OK. I only have the two WD 2TB drives and no data on there so I did the preclear in Screen (didn't do that before, as I understand it you don't need to do that on new drives?). Coming from Windows with limited *nix experience and doing this all from my Mac at home (which I am new to) all the command formats and the like are a little confusing. But I am getting there. It's pretty cool that you can get email notices on the status of the preclear script. Could take 24 hours or more from what I read. If that doesn't do it I might just bite the bullet and pick up a high quality Intel NIC from NewEgg (10/100/1000 card is $25 or so). Little confused on the Shares export / user shares principles so I will use my preclear down time to read up on that more I guess. I noticed something though and maybe this might shed light on this whole issue for someone that knows more about this than me (all of you, lol). I was copying large ISO files via UNC shares on windows and OSX. I copied some folders from my music collection, they went much quicker at first then stopped. I did it a few times to test and the interesting thing is that no matter what, only 64mb would copy. That's it. Not sure why that is though?
April 11, 201115 yr Author Attached is a screen cap from my email client showing the 2 disks. First stage took about 8 hours it looks like. Both are WD 2TB EARS drives and are both preclearing at the same time. That look pretty normal? Here is the detail of one disk after the pre read finished: Pre Read finished on /dev/sdc ( 2,000,388,096,000 of 2,000,398,934,016 bytes read) Pre Read Elapsed Time: 8:09:09 Total Elapsed Time: 8:09:10 Disk Temperature: 23C, Using Block size of 1,032,192 Bytes Calculated Read Speed - 68 MB/s
April 11, 201115 yr Attached is a screen cap from my email client showing the 2 disks. First stage took about 8 hours it looks like. Both are WD 2TB EARS drives and are both preclearing at the same time. That look pretty normal? yes, looks very normal.
April 12, 201115 yr I run preclear on every disk I am adding to my array at least once. Even brand new drives. The preclear can often find issues and possible problems with a brand new disk that may have issues or even be DOA. Better to let preclear find it for you. On average, preclearing a 2TB drive can run about 24-26 hrs. If you telnet in you can run multiple preclears... I wouldn't do more then four myself, but someone recently posted a pix running six at once... As for copying, are you copying to \\tower\share or to \\tower\diskx? May be related to your nic issues... Shawn
April 12, 201115 yr Author I was copying to \\tower\share I believe it showed up in Finder (on Mac) and in Explorer as: \flash \disk1 (there was an additional as parity, but I had parity not syncing) \TV Shows \Movies The last two were ones I added, obviously. I tried copying to \disk1 and the two shares. Silly to think about now, but was there ANY way the shares were added to the flash drive (there was no extra data there, however). Preclear is taking a long time. Sitting here now: Zeroing Disk /dev/sdc in progress: 50% complete. ( 1,000,224,063,488 of 2,000,398,934,016 bytes Wrote ) Disk Temperature: 22C, Next report at 75% Calculated Write Speed: 25 MB/s Elapsed Time of current cycle: 10:46:01 Total Elapsed time: 18:55:12 This is a good forum with helpful folks. Thanks.
June 1, 201115 yr Author I took some time away from this. Thought I would bring the thread back up. Still having issues. I was able to make shares and start moving information. I thought maybe it was cabling (loosely ran under the floor to the tower in another part of the house, a temporary thing). I have since moved the tower right into the home office (in all it's fan blowing glory). It's plugged into the same Apple Airport gigabit switch (swapped to some 10/100 switches in other attempts). I was able to preclear the two WD EARS drives (2tb, one as parity) and copy some files and larger ISO's (7gig). At BEST I would get 10mb write speeds... slowly dropping down to 3mb. I could actually get 3 or so drag and drop copies going at once, they hit about 2.5mb each but it would eventually become unavailable. This is from Mac and Win clients. The web interface remained available (it would slow down some). Ping times were < 1ms from multiple clients during the duration. Didn't witness a timeout. Eventually I said, f-it and started over. I am running preclear, from these steps verbatim including the email alerts. Two sessions are going at once. Here are some examples of the emails. It is still not finished (disc 1): Post Read in progress on /dev/sdb: 50% complete. ( 1,001,839,104,000 of 2,000,398,934,016 bytes read )at 54.8 MB/s Disk Temperature: 27C, Using Block size of 8,225,280 Bytes Next report at 75% Calculated Read Speed: 31 MB/s Elapsed Time of current cycle: 8:57:28 Total Elapsed time: 51:58:24 Zeroing Disk /dev/sdb Done. Zeroing Elapsed Time: 33:11:11 Total Elapsed Time: 42:59:49 Disk Temperature: 27C, Calculated Write Speed: 16 MB/s Some of the email/notices have Calculated Write Speeds of 60-80mb, but those are at different stages. Should it really take this long??
June 1, 201115 yr No, it shouldn't be that slow with the type of hardware you're using. I would have connected my parity drive to one of the onboard SATA ports. Parity is the choke-point for writes. Since you have unMenu installed, can you post the output of User Scripts->Disk Speed Test? Also, are you seeing many network errors? ifconfig eth0 or ethtool -S eth0
June 1, 201115 yr Author I am going to let pre-clear finish (good thing the 'rapture' came and went, otherwise I might be worried about time...). I *believe* I have it setup using the onboard satas at what I would call 0 (1,2,3) and then adding them on the cards ( believe it's 4 sata onboard, then it goes to the add-in cards). Maybe I should take out the addin cards and unplug the power on the carriages. The 500w would do it, I though (heck, only two drives are in there anyway). I have unmenu installed, I went through everything there and there were no red/error warnings. Scanned a log I dumped before pre-clear, looked ok (nothing glaring, but this is new to me). Network card looked OK. I can do the Disk Speed test, should I wait until preclear is done first? I did do that before I started pre-clear, the speeds were around 65mb if I remember correctly. Any other ideas? I will prob wait until pre-clear is done, then post the Disk Speed Test.
June 1, 201115 yr Author Maybe of note: pre-clear is running (it will be two days, still seems to be running) and I can get to http:\\tower but unmenu (:8080) times out... I can re-attach from my Mac/Terminal and get the screen session. So it doesn't 'die'... ================================================================== 1.9 = unRAID server Pre-Clear disk /dev/sdb = cycle 1 of 1, partition start on sector 64 = 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 = Post-Read in progress: 65% complete. ( 1,302,884,352,000 of 2,000,398,934,016 bytes read ) 47.9 MB/s Disk Temperature: 28C, Elapsed Time: 55:17:27
June 1, 201115 yr Yes, let it finish before doing other stuff. Board wisdom seems to say that two successful preclears should be mandatory. Sad (if cheap) state of affairs. BTW, I can't preclear more than ~3 drives at once (drive depending) without saturating the system bus, which means it takes forever. Now I preclear 3 max, get that array online and test the setup, then add other drives later. There's always another system in the house I can use as a preclear station.
Archived
This topic is now archived and is closed to further replies.