January 17, 201313 yr What are other users getting speed wise with that type of file size? It seems very, very slow to me. I'm using a cache drive (1 tb WD Black) on a gb network with cat6 cables. Both computers have 1000 lans so I know it's not the computers themselves. For copying the file, I'm on a Mac OSX Mountain Lion, using AFP. I tried using SMB and the speeds were even slower.
January 17, 201313 yr You should take a look at this thread http://lime-technology.com/forum/index.php?topic=22675.0 How much ram you have? People with more than 4GB ram have been getting real slow writes
January 17, 201313 yr Well, calculating that out, that looks like about 8MB/s. Which is common for 100Mb network or a really slow hard drive. I just did a test on each of the drives in my array. In Windows I see anywhere between 70MB/s and 90MB/s read speed depending on which drive I'm hitting, writing seems to be between 25MB/s and 40MB/s, but my slowest disks are also the most full, so I'm not writing to them very often. Strangely, my "Green" drives end up being faster than some of the 7200RPM drives, either that or I have some particularly slow (non-LP) Barracuda's, take your pick. I've had a few Mac's access mine, I don't remember it being particularly slow, but I wasn't copying a 60GB file. I do, however, do a nightly backup of my desktop to my UnRAID box and it's currently 87GB. The last restore I did didn't take much more than an hour, though, and not all of that time is copy-time (so that's at least 25MB/s, likely much faster.) I would double check that all the machines and switches are negotiating to Gb, though. It kind of sounds like something may have fallen back to 100Mb, or you have a very slow (or failing) WD Black drive. Maybe look at your current syslog in a web browser, http://[tower?]/log/syslog and see if there are any drive issues that SATA complains about in the logs, but aren't bad enough to Red-Ball a drive. They'll probably be the most recent entries at the bottom of the logs. I had some SATA reconnects last week due to either a bad cable or a drive not "agreeing" with the controller I put it on. Moved things around and all is well.
January 17, 201313 yr Author You should take a look at this thread http://lime-technology.com/forum/index.php?topic=22675.0 How much ram you have? People with more than 4GB ram have been getting real slow writes I only have 4 gb ram in the unraid machine. Well, calculating that out, that looks like about 8MB/s. Which is common for 100Mb network or a really slow hard drive. I just did a test on each of the drives in my array. In Windows I see anywhere between 70MB/s and 90MB/s read speed depending on which drive I'm hitting, writing seems to be between 25MB/s and 40MB/s, but my slowest disks are also the most full, so I'm not writing to them very often. Strangely, my "Green" drives end up being faster than some of the 7200RPM drives, either that or I have some particularly slow (non-LP) Barracuda's, take your pick. I've had a few Mac's access mine, I don't remember it being particularly slow, but I wasn't copying a 60GB file. I do, however, do a nightly backup of my desktop to my UnRAID box and it's currently 87GB. The last restore I did didn't take much more than an hour, though, and not all of that time is copy-time (so that's at least 25MB/s, likely much faster.) I would double check that all the machines and switches are negotiating to Gb, though. It kind of sounds like something may have fallen back to 100Mb, or you have a very slow (or failing) WD Black drive. Maybe look at your current syslog in a web browser, http://[tower?]/log/syslog and see if there are any drive issues that SATA complains about in the logs, but aren't bad enough to Red-Ball a drive. They'll probably be the most recent entries at the bottom of the logs. I had some SATA reconnects last week due to either a bad cable or a drive not "agreeing" with the controller I put it on. Moved things around and all is well. I'll double check everything and add a new machine to the mix tomorrow. There shouldn't be any reason it's not working properly though; especially, the wd black drive itself (passed 3 preclears with last one today and was put online today). Thanks for quick replies and things to look up.
January 17, 201313 yr Connect to the server with telnet or using the attached console, what does "ethtool eth0" and "ifconfig" show?
January 17, 201313 yr Better yet... login to the server thru telnet, start MC and do a disk to disk copy of a few large files, see what speeds are then.. If network is the bottleneck you will see fast speeds, if speeds are as slow then it is not the network. Also make sure you do not have other stuff running to get an as good as possible measurement.
January 17, 201313 yr Author I went on a windows laptop and my transfer rate is up to 24-28 MB/s. I'm not sure what the problem is but my mac was only putting out 100 mbiit even though it has a gb nic. Will try and figure out the hardware on my own since it's not a mac forum. Thank you for the help guys! I wish they had "reps" here so I could give some out edit// Well, was reading some more on the site and found this: Without a cache drive: unRAID 4.5.3 - average 20-30 MB/s, peak reported 40 MB/s* With a cache drive: unRAID 4.5.3 - average 50-60 MB/s, peak reported 101 MB/s* ***I'm on latest Unraid version*** So, obviously something is wrong Connect to the server with telnet or using the attached console, what does "ethtool eth0" and "ifconfig" show? Did this and it shows 10/100/1000 and all is working fine. I'm running mover manually and can't figure out how to see the speeds btwn the cache disks and other disks. edit 2// Well, server is back to doing 12 MB/s now. I don't get what is going since nothing has changed. The cache disc was freed up after mover ran and I'm still getting 12 MB/s. The 2 drives in the array are WD red 3 tb with no data on them before yesterday's move.
January 17, 201313 yr Did this and it shows 10/100/1000 and all is working fine. Those are potential speeds. What speed did it list as the current connection speed? As an example, on my server it says: ethtool eth0 Settings for eth0: Supported ports: [ TP ] 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: Twisted Pair PHYAD: 0 Transceiver: internal Auto-negotiation: on Supports Wake-on: umbg Wake-on: g Current message level: 0x00000007 (7) Link detected: yes
January 17, 201313 yr Author Did this and it shows 10/100/1000 and all is working fine. Those are potential speeds. What speed did it list as the current connection speed? As an example, on my server it says: ethtool eth0 Settings for eth0: Supported ports: [ TP ] 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: Twisted Pair PHYAD: 0 Transceiver: internal Auto-negotiation: on Supports Wake-on: umbg Wake-on: g Current message level: 0x00000007 (7) Link detected: yes Same as yours. 1000 Mb/s and Duplex: Full My setup isn't that far away from each other atm since I'm just trying to put everything on the server. The mac, the windows laptop and server are literally all within 3 feet of each other. The switch is behind the server and they are all plugged into that. It's an unmanaged Gb switch.
January 18, 201313 yr You should take a look at this thread http://lime-technology.com/forum/index.php?topic=22675.0 How much ram you have? People with more than 4GB ram have been getting real slow writes I thought that only affected specific hardware? I don't believe it affects everyone and every hardware type. This isn't CNN
January 18, 201313 yr Same as yours. 1000 Mb/s and Duplex: Full Great.... You confused me when you reported earlier "it shows 10/100/1000"
January 18, 201313 yr Author Same as yours. 1000 Mb/s and Duplex: Full Great.... You confused me when you reported earlier "it shows 10/100/1000" Ahh, sorry about that. Any tips on what to try next? I mean I changed all my cables to brand new cables i had lying around today in case it was a bad cable and no difference.
January 18, 201313 yr Bobbiemarker, It might be your switch, and it's probably worth taking a look at that. Don't assume that it's not at fault. I was using a Netgear 8 Gigabit port unmanaged switch plugged into my Billion 7800N (also Gigabit Ethernet), and the speeds I was getting was extremely slow from Win7 to my unraid server. After a bit of trouble shooting I took the switch out of the equation and plugged all the network cables into my router and the problem went away. I googled the switch and there were a number of others that had the same issue. I seemed that something went wrong with it but being unmanaged there is no way of checking anything.
January 18, 201313 yr Author Bobbiemarker, It might be your switch, and it's probably worth taking a look at that. Don't assume that it's not at fault. I was using a Netgear 8 Gigabit port unmanaged switch plugged into my Billion 7800N (also Gigabit Ethernet), and the speeds I was getting was extremely slow from Win7 to my unraid server. After a bit of trouble shooting I took the switch out of the equation and plugged all the network cables into my router and the problem went away. I googled the switch and there were a number of others that had the same issue. I seemed that something went wrong with it but being unmanaged there is no way of checking anything. Unfortunately, it's not the switch otherwise it would be an easy fix. I originally had it set up as follows: 1. mac to router & server to router = speeds of ~8 MB/s 2. I changed to win 7 laptop and did: laptop to router & server to router = speeds of 24-28 MB/s 3. I set up switch in room with all of them and did test of win 7 laptop to switch & server to switch = speeds of 24-28 MB/s then it went to complete shit and is at 11 MB/s currently. I did not change any hardware configs or anything within the /tower screen, or /tower:8080, besides manually doing the mover so the files on cache would go to the array. Not to mention switch is showing 1 gbps speed (the light on the switch shows what speed is current). My drives are as follows: parity: wd red 3 tb disk1: wd red 3 tb disk 2: wd red 3 tb cache drive: wd black 1 tb All those drives were blank and were brand new up until the day I did the preclears for them. All passed 3 preclears.
January 18, 201313 yr Author So looking around the cache thread I came upon this: I'm wondering if below stats are normal: "all my drives are WD green 3TB. i don't have cache drive setup. i'm interested in knowing if adding a cache drive will improve this stats significantly or not. /dev/sda: Timing cached reads: 11988 MB in 2.00 seconds = 6003.55 MB/sec Timing buffered disk reads: 56 MB in 3.07 seconds = 18.25 MB/sec /dev/sdb: Timing cached reads: 12502 MB in 2.00 seconds = 6260.80 MB/sec Timing buffered disk reads: 380 MB in 3.01 seconds = 126.14 MB/sec /dev/sdc: Timing cached reads: 12616 MB in 2.00 seconds = 6318.38 MB/sec Timing buffered disk reads: 372 MB in 3.01 seconds = 123.60 MB/sec /dev/sdd: Timing cached reads: 12444 MB in 2.00 seconds = 6231.35 MB/sec Timing buffered disk reads: 378 MB in 3.00 seconds = 125.86 MB/sec /dev/sde: Timing cached reads: 12482 MB in 2.00 seconds = 6251.29 MB/sec Timing buffered disk reads: 368 MB in 3.00 seconds = 122.64 MB/sec /dev/sdf: Timing cached reads: 12460 MB in 2.00 seconds = 6239.76 MB/sec Timing buffered disk reads: 362 MB in 3.01 seconds = 120.17 MB/sec /dev/sdg: Timing cached reads: 12446 MB in 2.00 seconds = 6233.11 MB/sec Timing buffered disk reads: 378 MB in 3.01 seconds = 125.51 MB/sec /dev/sdh: Timing cached reads: 12484 MB in 2.00 seconds = 6251.94 MB/sec Timing buffered disk reads: 376 MB in 3.01 seconds = 124.71 MB/sec" All of my drives are significantly lower than that when running the same hdparm command /dev/sda Timing cached reads: 3820 MB in 2.00 seconds = 1911.07 MB/sec Timing buffered disk reads: 444 MB in 3.01 seconds = 147.41 MB/sec /dev/sdb Timing cached reads: 3904 MB in 2.00 seconds = 1953.69 MB/sec Timing buffered disk reads: 428 MB in 3.01 seconds = 142.43 MB/sec /dev/sdc Timing cached reads: 3950 MB in 2.00 seconds = 1976.99 MB/sec Timing buffered disk reads: 428 MB in 3.01 seconds = 142.13 MB/sec Could this be my entire problem? I am transferring files to the cache drive at the moment however so I don't know if that will affect the numbers.
January 18, 201313 yr Not sure which are your drives, but all but the first drive looks very fast. The first one is about 18MB/s so is probably a flash drive, the next 7 drives are in the 120's so very fast drives, and the next 3 are in the 140's, really fast drives. For those timings, ignore all but the very last number. One possibility is network interference of some kind, or faulty router or switch. You may be connecting at gigabit speeds, but after too many packet issues, it's renegotiating down to 100mbps, resulting in about 11MB/s. After the slowdowns, run another ifconfig eth0 and check for RX and TX errors, dropped, overruns, and collisions.
January 18, 201313 yr Author Not sure which are your drives, but all but the first drive looks very fast. The first one is about 18MB/s so is probably a flash drive, the next 7 drives are in the 120's so very fast drives, and the next 3 are in the 140's, really fast drives. For those timings, ignore all but the very last number. One possibility is network interference of some kind, or faulty router or switch. You may be connecting at gigabit speeds, but after too many packet issues, it's renegotiating down to 100mbps, resulting in about 11MB/s. After the slowdowns, run another ifconfig eth0 and check for RX and TX errors, dropped, overruns, and collisions. The last 3 were my drives. The previous 10 or so were quotes from another thread. So, I guess I have very fast drives lol. I installed a Supermicro raid card and added some files and my speeds started out at 140 MB/s for both read and write and went down to 40-90. I just copied the same 2 gb file back and forth renaming it about 5 times and running the test about 5 times. Sometimes it would go down to 40 and other times it wouldn't go below 90. I'm not sure what would cause such a huge difference. I ran ifconfig again after the entire test and this is the result: for eth0: RX packets: 5792902 errors: 0 dropped: 663 overruns: 0 frame: 0 TX packets: 6284489 errors: 0 dropped: 0 overruns: 0 carrier: 0 Sorry, I'm not telnetted into it with the win laptop as the battery died so I'm just copying from the server screen after running the test. I looked at these number prior and there were never any errors, just dropped rx packets. I'm not sure what's causing but it might be the cable length to the switch. It's on a 75 foot cat 6 cable. The cables running from the laptop and server to the switch are cat 6 5 foot cables. The cables are brand new and the switch is brand new. The router is also newish (got it right after Hurricane Sandy wrecked my other stuff). edit//I just ran another ifconfig test and the dropped packets are up to 1162 now and I'm not even transferring files or using the server; it's just on so I'm assuming it's an app/add on in unmenu
January 18, 201313 yr Try setting the mem=4095m parameter from the thread link posted early. I only have 4gb or RAM, but setting the parameter improved my write speeds from 12-15 MB/s to a consistent 24-25 MB/s.
January 18, 201313 yr RX packets: 5792902 errors: 0 dropped: 663 overruns: 0 frame: 0 TX packets: 6284489 errors: 0 dropped: 0 overruns: 0 carrier: 0 edit//I just ran another ifconfig test and the dropped packets are up to 1162 now and I'm not even transferring files or using the server; it's just on so I'm assuming it's an app/add on in unmenu Were the collisions also zero? Except for the packet counts, those numbers should all be zeroes, only rarely should they be a bit above zero. If you have extra cables and time, I would swap them one by one, and monitor. Even brand new good quality cables may be defective. Also, check for power cables running extremely close to your network cables.
Archived
This topic is now archived and is closed to further replies.