Joe L. Posted February 15, 2011 Share Posted February 15, 2011 Thanks Guys...I feel better now. What hasn't been said (perhaps because it is too obvious) is that adding a 7200rpm cache drive should give you close to double those write speeds. with the trade-off being the files will not be protected by parity until they are subsequently moved to the protected array sometime in the middle of the night. The actual write to parity from the cache drive will still take as long and occur at the same rate. Quote Link to comment
johnodon Posted February 15, 2011 Share Posted February 15, 2011 with the trade-off being the files will not be protected by parity until they are subsequently moved to the protected array sometime in the middle of the night. The actual write to parity from the cache drive will still take as long and occur at the same rate. That is exactly what I do not like about the cache drive. To each his own...just not for me. Quote Link to comment
Giraffeninja Posted February 15, 2011 Share Posted February 15, 2011 I am really liking the speed of AFP, however it seems to be somewhat unreliable. It works fine if you connect then copy files right away, but it seems to "time out" after a while, even if it is in the middle of a transfer. I haven't been able to pin it down but it seems to be around 4 hours or so. I have resorted to SMB for everything but Time Machine, however I will switch back to AFP so I can get you a syslog. Joe, not that I'm complaining but I'm trying to figure out why my speeds are so far out of the normal. Setup: i3 iMac (10.6.6) -> 20 foot Ethernet -> Netgear Gigabit Switch -> 10 foot ethernet -> UnRaid 5b4 (i3 on Gigabyte H57M-USB3) Cache Samsung HD502HJ (7200rpm) all other drives Samsung HD154UI (5400rpm) 3 MKV files (10.81GB total) Here are the speeds I am getting: SMB No Cache SMB Cache AFP No Cache AFP Cache *All Data Disks and Parity are Samsung HD154UI (5400 rpm) Quote Link to comment
johnodon Posted February 15, 2011 Share Posted February 15, 2011 Just tried to create a new share are it did nothing. Never made it to the next section to set Security. Here is the snippet from teh syslog: Feb 15 13:55:57 unRAID emhttp: shcmd (41): rm /etc/samba/smb-shares.conf >/dev/null 2>&1 Feb 15 13:55:57 unRAID emhttp: shcmd (42): cp /etc/exports- /etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (43): echo '/mnt/disk1 -async,no_subtree_check,fsid=11 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (44): echo '/mnt/disk2 -async,no_subtree_check,fsid=12 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (45): echo '/mnt/user/Movies -async,no_subtree_check,fsid=100 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (46): echo '/mnt/user/Music -async,no_subtree_check,fsid=102 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (47): echo '/mnt/user/Public -async,no_subtree_check,fsid=105 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (48): echo '/mnt/user/Videos -async,no_subtree_check,fsid=101 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (49): echo '/mnt/user/XBMC -async,no_subtree_check,fsid=106 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (50): echo '/mnt/user/XBOX -async,no_subtree_check,fsid=103 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: Restart SMB... Feb 15 13:55:57 unRAID emhttp: shcmd (51): killall -HUP smbd Feb 15 13:55:57 unRAID emhttp: Restart NFS... Feb 15 13:55:57 unRAID emhttp: shcmd (52): exportfs -ra |logger Feb 15 13:55:57 unRAID emhttp: shcmd (53): /usr/local/sbin/emhttp_event svcs_restarted Feb 15 13:55:57 unRAID emhttp_event: svcs_restarted Quote Link to comment
Giraffeninja Posted February 15, 2011 Share Posted February 15, 2011 After looking a little further, I wonder if some of those write speeds being reported are processor capped. Using iStat to monitor CPU usage on my server AFP no Cache, SMB no Cache and SMB Cache all use around 8% of the CPU to Copy a file over. AFP Cache uses around 18%. Considering streaming a movie off the server doesn't even consume 1% (showing as 99-100% idle). I wonder if it's the Atom / Celeron servers reporting < 40MB/s. Quote Link to comment
johnodon Posted February 15, 2011 Share Posted February 15, 2011 Just tried to create a new share are it did nothing. Never made it to the next section to set Security. Here is the snippet from teh syslog: Feb 15 13:55:57 unRAID emhttp: shcmd (41): rm /etc/samba/smb-shares.conf >/dev/null 2>&1 Feb 15 13:55:57 unRAID emhttp: shcmd (42): cp /etc/exports- /etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (43): echo '/mnt/disk1 -async,no_subtree_check,fsid=11 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (44): echo '/mnt/disk2 -async,no_subtree_check,fsid=12 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (45): echo '/mnt/user/Movies -async,no_subtree_check,fsid=100 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (46): echo '/mnt/user/Music -async,no_subtree_check,fsid=102 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (47): echo '/mnt/user/Public -async,no_subtree_check,fsid=105 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (48): echo '/mnt/user/Videos -async,no_subtree_check,fsid=101 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (49): echo '/mnt/user/XBMC -async,no_subtree_check,fsid=106 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: shcmd (50): echo '/mnt/user/XBOX -async,no_subtree_check,fsid=103 *(rw,insecure,anongid=100,anonuid=99,all_squash)' >>/etc/exports Feb 15 13:55:57 unRAID emhttp: Restart SMB... Feb 15 13:55:57 unRAID emhttp: shcmd (51): killall -HUP smbd Feb 15 13:55:57 unRAID emhttp: Restart NFS... Feb 15 13:55:57 unRAID emhttp: shcmd (52): exportfs -ra |logger Feb 15 13:55:57 unRAID emhttp: shcmd (53): /usr/local/sbin/emhttp_event svcs_restarted Feb 15 13:55:57 unRAID emhttp_event: svcs_restarted Ignore. I tried from a different system and was able to create teh share without issue. This brings up anotehr point...is there a way to terminate your current session (in teh webgui)? I'm wondering if my session had just gone stale and I couldn't do anything. John Quote Link to comment
PeterB Posted February 16, 2011 Share Posted February 16, 2011 with the trade-off being the files will not be protected by parity until they are subsequently moved to the protected array sometime in the middle of the night. The actual write to parity from the cache drive will still take as long and occur at the same rate. True! ... and to each, their own. The way I use it is to perform my bulk copying or processing (and the cache drive halves the time I spend doing this) then, just before I walk away from the computer, I invoke mover. Later, when I come back to the computer, I delete the original files. Quote Link to comment
papnikol Posted February 16, 2011 Share Posted February 16, 2011 i've seen it suggested elsewhere, maybe it would be nice in the next version of unraid to have a mirroring of the cache drive. as a cache drive is usually small, it would not cost much to have 2 cache drives. actually, i think i have seen some people mentioning that they have 2 drives as cache using their mobos' raid functionality Quote Link to comment
lungnut Posted February 16, 2011 Share Posted February 16, 2011 The best performance is if you have 7200 RPM drives in both the parity disk AND for the data disk being written. They have been measured at speeds close to 40MB/s on a highly optimized array. I personally get somewhere around 35 MB/s with all 7200 RPM drives involved. Maybe I'm asking for a blow to my confidence in my own build, but I've been thinking that I've been getting sustained writes to my 7200RPM parity drive of about 60-70MB/s (see my pic). I am not using a beta version (but I check the boards fairly regularly for upcoming features, build reports, etc.). Have I been misunderstanding the speeds I've been getting? I can regularly move 3 to 4 GB's worth of files to my server in about a minute. Sorry for hijacking the thread... Quote Link to comment
Joe L. Posted February 16, 2011 Share Posted February 16, 2011 The best performance is if you have 7200 RPM drives in both the parity disk AND for the data disk being written. They have been measured at speeds close to 40MB/s on a highly optimized array. I personally get somewhere around 35 MB/s with all 7200 RPM drives involved. Maybe I'm asking for a blow to my confidence in my own build, but I've been thinking that I've been getting sustained writes to my 7200RPM parity drive of about 60-70MB/s (see my pic). I am not using a beta version (but I check the boards fairly regularly for upcoming features, build reports, etc.). Have I been misunderstanding the speeds I've been getting? I can regularly move 3 to 4 GB's worth of files to my server in about a minute. Sorry for hijacking the thread... you are probably writing to your RAM and it gives an initial boost to the apparent "sustained" rate until it cannot buffer any more. I'll bet you have at least 4Gig of RAM. How much memory do you have? Joe L. Quote Link to comment
lungnut Posted February 16, 2011 Share Posted February 16, 2011 The best performance is if you have 7200 RPM drives in both the parity disk AND for the data disk being written. They have been measured at speeds close to 40MB/s on a highly optimized array. I personally get somewhere around 35 MB/s with all 7200 RPM drives involved. Maybe I'm asking for a blow to my confidence in my own build, but I've been thinking that I've been getting sustained writes to my 7200RPM parity drive of about 60-70MB/s (see my pic). I am not using a beta version (but I check the boards fairly regularly for upcoming features, build reports, etc.). Have I been misunderstanding the speeds I've been getting? I can regularly move 3 to 4 GB's worth of files to my server in about a minute. Sorry for hijacking the thread... you are probably writing to your RAM and it gives an initial boost to the apparent "sustained" rate until it cannot buffer any more. I'll bet you have at least 4Gig of RAM. How much memory do you have? Joe L. I have 2GB of RAM in my Unraid system, and I consistently get those writes. No bursts. I can transfer 100GB worth of VIDEO_TS folders in less than 35 minutes, easily. I do it all the time. Quote Link to comment
limetech Posted February 16, 2011 Author Share Posted February 16, 2011 Found a small bug...not sure if this was mentined before. If a filename has an ' in it, the size will be listed as 0 in the webgui: John Thanks for pointing this out - fixed in 5.0-beta5. I long ago banished all special chars from file names because of issues with various players, utilities, etc. Quote Link to comment
Joe L. Posted February 16, 2011 Share Posted February 16, 2011 The best performance is if you have 7200 RPM drives in both the parity disk AND for the data disk being written. They have been measured at speeds close to 40MB/s on a highly optimized array. I personally get somewhere around 35 MB/s with all 7200 RPM drives involved. Maybe I'm asking for a blow to my confidence in my own build, but I've been thinking that I've been getting sustained writes to my 7200RPM parity drive of about 60-70MB/s (see my pic). I am not using a beta version (but I check the boards fairly regularly for upcoming features, build reports, etc.). Have I been misunderstanding the speeds I've been getting? I can regularly move 3 to 4 GB's worth of files to my server in about a minute. Sorry for hijacking the thread... you are probably writing to your RAM and it gives an initial boost to the apparent "sustained" rate until it cannot buffer any more. I'll bet you have at least 4Gig of RAM. How much memory do you have? Joe L. I have 2GB of RAM in my Unraid system, and I consistently get those writes. No bursts. I can transfer 100GB worth of VIDEO_TS folders in less than 35 minutes, easily. I do it all the time. Please post your hardware and disk config. I'm sure many others would love to have the same speeds to parity protected disks. Quote Link to comment
papnikol Posted February 16, 2011 Share Posted February 16, 2011 The best performance is if you have 7200 RPM drives in both the parity disk AND for the data disk being written. They have been measured at speeds close to 40MB/s on a highly optimized array. I personally get somewhere around 35 MB/s with all 7200 RPM drives involved. Maybe I'm asking for a blow to my confidence in my own build, but I've been thinking that I've been getting sustained writes to my 7200RPM parity drive of about 60-70MB/s (see my pic). I am not using a beta version (but I check the boards fairly regularly for upcoming features, build reports, etc.). Have I been misunderstanding the speeds I've been getting? I can regularly move 3 to 4 GB's worth of files to my server in about a minute. Sorry for hijacking the thread... you are probably writing to your RAM and it gives an initial boost to the apparent "sustained" rate until it cannot buffer any more. I'll bet you have at least 4Gig of RAM. How much memory do you have? Joe L. I have 2GB of RAM in my Unraid system, and I consistently get those writes. No bursts. I can transfer 100GB worth of VIDEO_TS folders in less than 35 minutes, easily. I do it all the time. This is about 50MB/s and it is almost what i get in my pc internally. it doesn't seem right. the best i get with 7200rpm HDDs in unraid is 35MB/s. (i have sata I in my mobo, but i do not think this is the limiting factor). could it be that your parity is not working? Or, maybe, you have an amazing machine... What are your specs? Quote Link to comment
lungnut Posted February 16, 2011 Share Posted February 16, 2011 The best performance is if you have 7200 RPM drives in both the parity disk AND for the data disk being written. They have been measured at speeds close to 40MB/s on a highly optimized array. I personally get somewhere around 35 MB/s with all 7200 RPM drives involved. Maybe I'm asking for a blow to my confidence in my own build, but I've been thinking that I've been getting sustained writes to my 7200RPM parity drive of about 60-70MB/s (see my pic). I am not using a beta version (but I check the boards fairly regularly for upcoming features, build reports, etc.). Have I been misunderstanding the speeds I've been getting? I can regularly move 3 to 4 GB's worth of files to my server in about a minute. Sorry for hijacking the thread... you are probably writing to your RAM and it gives an initial boost to the apparent "sustained" rate until it cannot buffer any more. I'll bet you have at least 4Gig of RAM. How much memory do you have? Joe L. I have 2GB of RAM in my Unraid system, and I consistently get those writes. No bursts. I can transfer 100GB worth of VIDEO_TS folders in less than 35 minutes, easily. I do it all the time. Please post your hardware and disk config. I'm sure many others would love to have the same speeds to parity protected disks. Geez, I'm an idiot. I used the word "parity" when I meant "cache." I think you suspected my mistake and you were kind enough to phrase your answer in very hard-to-misunderstand language. So, in short, I should have written: "Maybe I'm asking for a blow to my confidence in my own build, but I've been thinking that I've been getting sustained writes to my 7200RPM CACHE drive of about 60-70MB/s (see my pic)." Quote Link to comment
lungnut Posted February 16, 2011 Share Posted February 16, 2011 could it be that your parity is not working? More like "my brain isn't working." See my clarification just above. (I meant 70MB/s writes to my cache drive, not parity). Quote Link to comment
johnodon Posted February 16, 2011 Share Posted February 16, 2011 Man...I was about to throw my whole system out the window! Quote Link to comment
CHBMB Posted February 16, 2011 Share Posted February 16, 2011 As I am beginner to unraid it is unwise to use beta 4 on my production server? I agree, it would be a bad idea, especially for a beginner, especially since it has only been out for a few days. 4.7 is a MUCH better version for your production use at this time. Most of us doing the testing have 5.0b4 on a second server, one used for testing. I'm using 5.0b4 on my sole setup, BUT do have all my data backed up, so if it all went horribly wrong I can just recopy 6TB of data to a fresh array. Having said that it seems to work fine and as long as you're careful I think it's ok, oddly I'd never in a million years dream of using MS beta software in the same fashion. I'm still having a problem with permissions though in this version, the new script works fine, but if I create a directory in my cache drive although it will show up in my shares I cannot access it from two of my PCs. Rerunning the permissions script fixes the issue though. It may be the way I'm using the cache drive though and I'm not sure if I've tried copying directly to the user share which would in effect just do what I'm doing anyway automatically. I'm loving the new 5.0 UnRAID though, truly wonderful. Just got to get as sleep script working, install the final version of 5.0 when it's released and then leave well alone. Neil Still having some problems with permissions, create folders on the server, copy data to it via a cache drive and then can only access it on my main PC, not my laptops or HTPC. I don't have any security settings on any shares at all. Basically wide open full access to anyone on my home network. Quote Link to comment
SSD Posted February 18, 2011 Share Posted February 18, 2011 I bought a BR10i and have been running some experiments with the controller and 5.0b4. My hope was to be able to configure the controller to provide a RAID-1 array that I could use for cache, and 6 JBOD disks. My tests so far have been with 2 disks (there is a third disk on the motherboard controller). I booted unRAID 5.0b4 and the 2 disks were seen. UnRAID did not display their temperature or spindown status, but I was able to get the temperature and spindown the disks manually using "hdparm -y" and then see their spindown status using "hdparm -C". Looks like it if very close to working. I then rebooted and configured the 2 disks to be a RAID-1 array. unRAID saw this disk and allowed me to assign it to the array. The name was somewhat weird ("-3600508e000000000cc0e922e8ac34209"), but it works!. But I am not able to spin it down. In looking at the syslog, it looks like the driver did a good job of figuring out what was going on. It lists the drives INSIDE the array. I am posting my final syslog with the RAID-1 array assigned. Hope it is helpful. syslog.br10i.txt Quote Link to comment
queeg Posted February 19, 2011 Share Posted February 19, 2011 I'm in the process of testing a MSI 880GM-E43 motherboard for level 1 and I put a BR10i adapter in for the test. 40MB throughput. Edit: At 50% done, I'm seeing one of the disks (500GB) is appearing as spun down in the unraid web page and the green ball is blinking. All these drives are on the BR10i and I thought spin down wasn't supposed to be working yet? Quote Link to comment
Joe L. Posted February 19, 2011 Share Posted February 19, 2011 Edit: At 50% done, I'm seeing one of the disks (500GB) is appearing as spun down in the unraid web page and the green ball is blinking. All these drives are on the BR10i and I thought spin down wasn't supposed to be working yet? Many drives support their own internal spin-down timers. Quote Link to comment
queeg Posted February 19, 2011 Share Posted February 19, 2011 Edit: At 50% done, I'm seeing one of the disks (500GB) is appearing as spun down in the unraid web page and the green ball is blinking. All these drives are on the BR10i and I thought spin down wasn't supposed to be working yet? Many drives support their own internal spin-down timers. I checked again and unRAID is spinning down the 500GB drive. To test it further I set unraid to spin down after 15 minutes and started the parity check. I watched as the check got to the end of the small drive and 15 min later it spun down. So then I set the spin down time to 30 min and cd to /mnt/disk1 to get the drive to spin up again. Then exactly 30 min later it spun down again. Edit: After the parity check finished I clicked the spin down button on unRAID main page and all drives spun down immediately. Quote Link to comment
coppit Posted February 20, 2011 Share Posted February 20, 2011 I'm not sure if this is related to the upgrade, but my Windows XP machine can't resolve my server by name any more. My mac machine can. I rebooted my Windows machine and also tried setting local master on the server. Any ideas? Quote Link to comment
johnodon Posted February 20, 2011 Share Posted February 20, 2011 I'm not sure if this is related to the upgrade, but my Windows XP machine can't resolve my server by name any more. My mac machine can. I rebooted my Windows machine and also tried setting local master on the server. Any ideas? I have had this problem after almost every rebuild. I need to connect by IP. Eventually I will be able to connect to the name, but not right away. Must have something to do with DNS I would think. John Quote Link to comment
Joe L. Posted February 20, 2011 Share Posted February 20, 2011 I'm not sure if this is related to the upgrade, but my Windows XP machine can't resolve my server by name any more. My mac machine can. I rebooted my Windows machine and also tried setting local master on the server. Any ideas? I have had this problem after almost every rebuild. I need to connect by IP. Eventually I will be able to connect to the name, but not right away. Must have something to do with DNS I would think. John What do you have your DNS server set to on your Win-XP box? It would normally be set to the IP address of your router. The router, in turn, should have its DNS server name set to your ISPs domain name server (Or some other DNS server you might prefer) Frequently, windows ends up being set by your ISP/browser to your ISP's domain name server. Then all requests for name translation go there first, and then possibly to your local lan. (if they do not re-route it to one of their "help" pages) Joe L. Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.