jus7incase Posted June 28, 2013 Share Posted June 28, 2013 Hi there, on my OSX10.8 machines mounting an AFP share in Finder takes really long (up to a minute!) What is wrong? Thanks for caring, JC Quote Link to comment
madburg Posted June 28, 2013 Share Posted June 28, 2013 How many data drives (and if you have parity and cache drives) installed? Quote Link to comment
adriencater Posted June 29, 2013 Share Posted June 29, 2013 I've been having this as well. Directory read can also take a while, and my TimeMachine backups often get corrupted. rc15a (but was acting up on previous install of rc12). running 4 drives, 3GBs each, and a 1GB drive for TM, one parity, no cache. Quote Link to comment
madburg Posted June 29, 2013 Share Posted June 29, 2013 I've been having this as well. Directory read can also take a while, and my TimeMachine backups often get corrupted. rc15a (but was acting up on previous install of rc12). running 4 drives, 3GBs each, and a 1GB drive for TM, one parity, no cache. Well, ju7incase, didn't answer, so to you adriencater, I assume u meant (4) 3TB drive not 3 GB drives and u allocated 1GB for a TM? Run the following command and post here: df -aTh And while you are at it check what extended attribute you have (since you don't have a cache drive) run the following command: (It will take awhile if you have a lot of files, running out now, so be back later.) getfattr -R -d /mnt/user0/ Quote Link to comment
adriencater Posted June 30, 2013 Share Posted June 30, 2013 From the admin panel: Device Identification Temp. Size Free Reads Writes Errors parity WDC_WD30EZRX-00MMMB0_WD-WCAWZ2453611 (sde) 2930266532 39°C 3 TB - 350938 375361 0 disk1 WDC_WD30EZRX-00MMMB0_WD-WCAWZ2456458 (sdd) 2930266532 33°C 3 TB 137.15 GB 15947 111 0 disk2 WDC_WD30EZRX-00MMMB0_WD-WCAWZ2456364 (sdb) 2930266532 33°C 3 TB 1.3 TB 174 96 0 disk3 ST3000DM001-9YN166_Z1F15JTV (sdc) 2930266532 33°C 3 TB 143.65 GB 116 67 0 disk4 WDC_WD30EZRX-00DC0B0_WD-WMC1T0037466 (sdf) 2930266532 35°C 3 TB 2.3 TB 76 73 0 disk23 WDC_WD10EADS-00L5B1_WD-WCAU45627201 (sdg) 976762552 36°C 1 TB 308.28 GB 1655457 453326 0 flash Reader_SD_MS - 3.97 GB 3.87 GB 330 11 0 and from "df -aTh": Filesystem Type Size Used Avail Use% Mounted on proc proc 0 0 0 - /proc sysfs sysfs 0 0 0 - /sys tmpfs tmpfs 128M 188K 128M 1% /var/log /dev/sda1 vfat 3.7G 79M 3.7G 3% /boot /dev/md1 reiserfs 2.8T 2.7T 128G 96% /mnt/disk1 /dev/md2 reiserfs 2.8T 1.6T 1.2T 57% /mnt/disk2 /dev/md3 reiserfs 2.8T 2.6T 134G 96% /mnt/disk3 /dev/md4 reiserfs 2.8T 650G 2.1T 24% /mnt/disk4 /dev/md23 reiserfs 932G 645G 287G 70% /mnt/disk23 shfs fuse.shfs 12T 8.1T 3.9T 68% /mnt/user and, yes, apologies, those are four 3TB drives –– not GB, it's not 1999 anymore! -- plus one 1TB drive which is devoted to timemachine. thanks Quote Link to comment
madburg Posted July 1, 2013 Share Posted July 1, 2013 /dev/md1 reiserfs 2.8T 2.7T 128G 96% /mnt/disk1 /dev/md3 reiserfs 2.8T 2.6T 134G 96% /mnt/disk3 Disk 1 & 3 are over 95% full, you need to move data off of them to another disk(s) to free up space so they are no more than 95% full (at least 150GB of free space). While this sucks, I noticed with AFP, connection times to unRAID are all messed up as well as performance and other old ball behaviour. So free up the space re-start the array and report back on the connection times. So freeing up disk space on those 2 drives is the first thing you want to do. And while you are at it check what extended attribute you have (since you don't have a cache drive) run the following command: (It will take awhile if you have a lot of files, running out now, so be back later.) getfattr -R -d /mnt/user0/ Don't forget to check for corrupted extended attributes with the command above. Also, where is your netatalk db being stored? (there are different sh&tty parts to all this and each has its own effect) Quote Link to comment
adriencater Posted July 1, 2013 Share Posted July 1, 2013 Disk 1 & 3 are over 95% full, you need to move data off of them to another disk(s) to free up space so they are no more than 95% full (at least 150GB of free space). While this sucks, I noticed with AFP, connection times to unRAID are all messed up as well as performance and other old ball behaviour. So free up the space re-start the array and report back on the connection times. So freeing up disk space on those 2 drives is the first thing you want to do. This seems very logical. I'll give that a go, and see how things work out. And while you are at it check what extended attribute you have (since you don't have a cache drive) run the following command: (It will take awhile if you have a lot of files, running out now, so be back later.) getfattr -R -d /mnt/user0/ Don't forget to check for corrupted extended attributes with the command above. Also, where is your netatalk db being stored? (there are different sh&tty parts to all this and each has its own effect) I'll look into these issues as well. I haven't been getting any "CNID DB" error messages, but this gives me a good hook to search for further information here and elsewhere – thanks! Quote Link to comment
jus7incase Posted July 1, 2013 Author Share Posted July 1, 2013 How many data drives (and if you have parity and cache drives) installed? Device Identification Temp. Size Free Reads Writes Errors parity ST3000DM001-1CH166_Z1F2CRHY (sdb) 2930266532 28°C 3 TB - 55736 57858 0 disk1 [browse /mnt/disk1] WDC_WD30EZRX-00AZ6B0_WD-WCC070288205 (sdc) 2930266532 29°C 3 TB 886.28 MB 147934 6384 0 disk2 [browse /mnt/disk2] WDC_WD30EZRX-00AZ6B0_WD-WCC070304485 (sdd) 2930266532 30°C 3 TB 1.25 GB 1378704 44312 0 disk3 [browse /mnt/disk3] ST3000DM001-1CH166_Z1F2G94V (sdf) 2930266532 27°C 3 TB 2.23 TB 15915 16754 0 cache [browse /mnt/cache] SAMSUNG_SP2504C_S09QJ1SL904777 (sde) 244198552 33°C 250.06 GB 180.8 GB 178248 667718 0 It is not the time that the drives take to spin up. I took over 2 minutes today. After that everything was in ok speed. I spun the drives down and did the same. Speed was ok. I also had a look with Wireshark, but everyting looked good. Next time it takes long I will check with Wireshark again, if anything can be seen there. Quote Link to comment
adriencater Posted July 1, 2013 Share Posted July 1, 2013 And while you are at it check what extended attribute you have (since you don't have a cache drive) run the following command: (It will take awhile if you have a lot of files, running out now, so be back later.) getfattr -R -d /mnt/user0/ I ran this command, and it generated ~7k lines of output (about 2400 individual items) before the server sneezed a bunch of lines ending in "Transport endpoint is not connected" and the whole thing stopped These are the types of extended attributes that were found: user.com.apple.diskimages.fsck user.com.apple.diskimages.recentcksum user.com.apple.metadata:kMDItemFinderComment user.com.apple.metadata:kMDItemWhereFroms user.com.apple.metadata:kMDLabel_… user.com.apple.Preview.UIstate.v1 user.com.apple.quarantine user.com.apple.TextEncoding user.com.apple.xcode.PlistType user.com.macromates.caret user.org.netatalk.supports-eas.G7qPCu Quote Link to comment
madburg Posted July 1, 2013 Share Posted July 1, 2013 And while you are at it check what extended attribute you have (since you don't have a cache drive) run the following command: (It will take awhile if you have a lot of files, running out now, so be back later.) getfattr -R -d /mnt/user0/ I ran this command, and it generated ~7k lines of output (about 2400 individual items) before the server sneezed a bunch of lines ending in "Transport endpoint is not connected" and the whole thing stopped These are the types of extended attributes that were found: user.com.apple.diskimages.fsck user.com.apple.diskimages.recentcksum user.com.apple.metadata:kMDItemFinderComment user.com.apple.metadata:kMDItemWhereFroms user.com.apple.metadata:kMDLabel_… user.com.apple.Preview.UIstate.v1 user.com.apple.quarantine user.com.apple.TextEncoding user.com.apple.xcode.PlistType user.com.macromates.caret user.org.netatalk.supports-eas.G7qPCu You need to run that command from the unraid console (or telnet/ssh, etc...), not from a client. You are looking out for user.org.netatalk.supports-eas.xxxxx : Exec format errror P.S. I doubt you will see anything of use in wireshark trace. The issue(s) are with unRAID and its AFP/netatalk, I gave the big three that cause issues above (drive space, corrupted extended attributes and netatalk DB). Tom stated he would look into updating some parts after 5.0 went finally. So at the moment those are the things you can do, outside of that its a waiting game unfortunately. Quote Link to comment
adriencater Posted July 1, 2013 Share Posted July 1, 2013 You need to run that command from the unraid console (or telnet/ssh, etc...), not from a client. You are looking out for user.org.netatalk.supports-eas.xxxxx : Exec format errror i was running it on a telnet session on the tower... everything keeps pooping out with a "Transport endpoint is not connected" error. also working on that issue over here: http://lime-technology.com/forum/index.php?topic=28282.0 I'll reboot, run it again, and keep an eye open for any user.org.netatalk.supports-eas.xxxxx : Exec format errror thanks for the help! Quote Link to comment
madburg Posted July 1, 2013 Share Posted July 1, 2013 You need to run that command from the unraid console (or telnet/ssh, etc...), not from a client. You are looking out for user.org.netatalk.supports-eas.xxxxx : Exec format errror i was running it on a telnet session on the tower... everything keeps pooping out with a "Transport endpoint is not connected" error. also working on that issue over here: http://lime-technology.com/forum/index.php?topic=28282.0 I'll reboot, run it again, and keep an eye open for any user.org.netatalk.supports-eas.xxxxx : Exec format errror thanks for the help! Sorry I assumed since you received those "Transport endpoint is not connected" (client trying to connect to the server) in the syslog you were running the command from a client. A recommendation, shut down all your clients so nothing is connecting to the unRAID server and rerun the command from the console, pipe it to a file would be the way to go and this way you could review it offline and search for those particular extended attributes easily, as trying to copy/paste it out to a log file is a bit of work from a command window (clearly you use macs to perform work, thus the large amounts of extended attributes on your array, understandingly. Some (I for one) only use mac clients to read from the array (for the most part)). Come to think of it in your case you will probably want to check none of your extended attributes are corrupted, based on the releases you have run. Best guess is any of them would carry the word "error" at the end. Quote Link to comment
jus7incase Posted July 22, 2013 Author Share Posted July 22, 2013 I ran getfattr -R -d /mnt/user0/ via ssh and it showed (no single error!): getfattr: Removing leading '/' from absolute path names user.org.netatalk.supports-eas.vgivaP=0seWVzAA== user.org.netatalk.supports-eas.pGYyAU=0seWVzAA== user.org.netatalk.supports-eas.0S8O5d=0seWVzAA== user.org.netatalk.supports-eas.Wmipmo=0seWVzAA== user.org.netatalk.supports-eas.gEgkH8=0seWVzAA== user.org.netatalk.supports-eas.rXaSSP=0seWVzAA== user.org.netatalk.supports-eas.rho0PU=0seWVzAA== user.com.apple.quarantine="0006;5174f02a;iPhoto;" user.com.apple.metadata:com_apple_backup_excludeItem=0sYnBsaXN0MDBfEBFjb20uYXBwbGUuYmFja3VwZAgAAAAAAAABAQAAAAAAAAABAAAAAAAAAAAAAAAAAAAAHA== user.com.apple.metadata:com_apple_backup_excludeItem=0sYnBsaXN0MDBfEBFjb20uYXBwbGUuYmFja3VwZAgAAAAAAAABAQAAAAAAAAABAAAAAAAAAAAAAAAAAAAAHA== user.com.apple.quarantine="0006;5174f029;iPhoto;" user.com.apple.quarantine="0006;5174f029;iPhoto;" user.com.apple.quarantine="0006;5174f028;iPhoto;" user.com.apple.quarantine="0002;51ac9d72;QuickTime Player;" user.com.apple.quarantine="0006;5174f027;iPhoto;" user.com.apple.quarantine="0006;5174f027;iPhoto;" user.org.netatalk.supports-eas.T12lp8=0seWVzAA== user.org.netatalk.supports-eas.YGekvc=0seWVzAA== user.org.netatalk.supports-eas.h6NxpC=0seWVzAA== user.org.netatalk.supports-eas.zBPWQ1=0seWVzAA== user.com.apple.metadata:kMDLabel_26x2uentpjgt7lka65qdcazuya=0sYnBsaXN0MDAzQbdtN9YjC/IIAAAAAAAAAQEAAAAAAAAAAQAAAAAAAAAAAAAAAAAAABE= user.com.apple.metadata:kMDLabel_26x2uentpjgt7lka65qdcazuya=0sYnBsaXN0MDAzQbdtNtlJ3EAIAAAAAAAAAQEAAAAAAAAAAQAAAAAAAAAAAAAAAAAAABE= user.com.apple.metadata:kMDLabel_26x2uentpjgt7lka65qdcazuya=0sYnBsaXN0MDAzQbcT4KUTG2kIAAAAAAAAAQEAAAAAAAAAAQAAAAAAAAAAAAAAAAAAABE= ...more labeled files.... user.com.apple.quarantine="0006;51674cc9;TextEdit;" user.com.apple.quarantine="0006;51674cc9;TextEdit;" still, AFP is extremely slow to mount and list directories... unRAID5 rc15a Quote Link to comment
defected07 Posted July 22, 2013 Share Posted July 22, 2013 I wanna play, too. I've had this problem with AFP for as long as I can remember -- possibly prior the 5.0 betas? Anyways, here's df -aTh Filesystem Type Size Used Avail Use% Mounted on proc proc 0 0 0 - /proc sysfs sysfs 0 0 0 - /sys tmpfs tmpfs 128M 8.5M 120M 7% /var/log /dev/sda1 vfat 3.8G 637M 3.2G 17% /boot /dev/md1 reiserfs 1.9T 1.8T 38G 98% /mnt/disk1 /dev/md2 reiserfs 2.8T 2.4T 349G 88% /mnt/disk2 /dev/md3 reiserfs 1.9T 1012G 852G 55% /mnt/disk3 /dev/md4 reiserfs 1.4T 1.4T 31G 98% /mnt/disk4 /dev/md5 reiserfs 1.9T 1.3T 612G 68% /mnt/disk5 /dev/sdb1 reiserfs 466G 22G 445G 5% /mnt/cache shfs fuse.shfs 9.6T 7.8T 1.9T 81% /mnt/user0 shfs fuse.shfs 11T 7.8T 2.3T 78% /mnt/user I only received 2 results when running the getfattr command you supplied, and it was right in the beginning: # file: mnt/user0/Media user.org.netatalk.supports-eas.7NFe3q user.org.netatalk.supports-eas.CG9zsS So, besides making sure all of my disks are less than 95% used, is there anything else to check out? Quote Link to comment
madburg Posted July 22, 2013 Share Posted July 22, 2013 @defected07, @jus7incase Read the following and run Tom's script http://lime-technology.com/forum/index.php?topic=28484.msg254158#msg254158 Quote Link to comment
jus7incase Posted July 24, 2013 Author Share Posted July 24, 2013 @defected07, @jus7incase Read the following and run Tom's script http://lime-technology.com/forum/index.php?topic=28484.msg254158#msg254158 Hi madburg, I followed the instructions and applied toms script to all /mnt/disk?/* where there were attributes. All attributes on the top level folders got removed. It did _not_ improve mount times via AFP whatsoever. Quote Link to comment
defected07 Posted July 24, 2013 Share Posted July 24, 2013 I ran the script, the attributes went away, now I checked disk1 today and it's back on the same folder ... 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.