April 3, 201412 yr would it be possible to point unmenu's mysql at this disk as well? I'm thinking I want to try this out tonight for plex, but it will involve moving my current array's disk 20 out of the array as that is my SSD. Catch is I also have mysql pointed to disk20. I'd hope/imagine it will just be a matter of updating the "disk label" field in unmenu's package manager. Since this isn't going to be an unraid drive anymore I wasn't sure if it would be as simple as updating it to say appdisk or if it would require a more specific path, or if it just wouldn't work at all and I'll have to look into manually setting up mysql which is a bit more daunting as I'm just learning how to play in linux here.
April 3, 201412 yr Author errmmmm i think that is more of an unmenu or new mysql type question. I mean you are able to choose whatever mount name you want if that matter for your situation.
April 3, 201412 yr Yeah I figured as much, but decided it was worth a shot in asking in case someone had already used your steps and started putting other apps like this on the drive. I'll just give it a shot tonight and if it doesn't work start desperately seeking an unmenu/mysql/non array disk knowledgeable person... eep. Worse case sql will go back on an array protected disk since it isn't as unruly as plex with the constant drive spinning.
April 3, 201412 yr Author I guess I'm just not sure what your question / concern is. Anything that is currently hosted on the array can (should) be able to be housed instead on a non-array drive. You just have to make sure you set paths correctly. I know when I dabbled with both virtualbox and then xen I was storing my VM's and ISO's on my appdisk and it worked fine.
April 3, 201412 yr why are you using mysql outside of your VM? Purpose of xen and the VM containers is to isolate every extra from unRAID.... Myk
April 4, 201412 yr My question/concern literally was just around the setting of the path correctly. This is pretty much my first time playing with drives like this outside of windows and I'm learning as I go. I went ahead and tried to use appdisk for my path and it did indeed work. Unfortunately now I'm getting random freezes where I can't even type on the command prompt directly on the unraid console. Decided to call it a night for now and let it get through its parity check from removing the SSD since it doesn't seem to lock up if I just let it sit. I've a feeling some config file somewhere is still referencing "disk20" which no longer exist. I'll (hopefully) sort it out this weekend.
April 4, 201412 yr Author so you think you might have a plug-in referencing disk 20? Seems easy enough to track down, but I don't know how complex your server is. In any case, good luck and let us know if you run into any sticky wickets. Anyone checking this thread might have similar concerns even if the final solution is to be found in another topic.
April 4, 201412 yr Yeah I hope to sort it out this weekend, but first I want to let it finish its parity check which it should do in about 6 more hours. After sleeping on it I started to wonder if perhaps my SSD doesn't support some of the features used in the mount command (noatime,nodiratime,discard). I'm pretty sure I updated the firmware to support TRIM, but I'll have to double check as this is one of the old original OCZ Vertex drives. I don't know about the other two and if they need any special support for them on the disk itself. It has run fine for about six hours just doing a parity check so that does seem to confirm the issue happens when I try and access this disk (its been happening as I try and reconfigure plex and/or mysql). As it is I think my next course of action is going to be to wipe it, mysql, and plex out entirely and start from scratch with them, but not until unraid is back to happy with its parity. Fortunately my server is not very complicated as I'm just getting into it. This drive and the apps I'm trying to run off it (plex and mysql) are the only thing special. I had it working perfectly hosting my xbmc media and database, but I just had to try and play with plex...
April 4, 201412 yr Author Well other than trim those other "features" are just instructions to the OS telling it to not update the last access times of files and directories. But in either case, I'm running an OCZ Vertex 2 with those exact parms. As to wiping that drive ... um ... [slaps forehead] ... you pulled this SSD from the array? So it is still formatted as ReisferFS? If so that is *A* problem. RFS doesn't support trim ... as best I was ever able to find at least. I don't know what happens when you try to mount with discard on an FS that doesn't support it, but at best it is ineffective, at worst it causes actual problems like freezes. Now that said, if your plex library and MySQL db are pretty large, you can easily copy them back to the array (DOH!) format your SSD ext4, then copy them back to avoid having to regenerate. I know my plex library would take days to recreate because of all the index files. If you don't have those then I suppose it wouldn't be too bad.
April 4, 201412 yr oh no no, I have completely reformatted it using the directions from the beginning of this thread no worries there. A complete rebuild of the database would not be that bad it is just for my xbmc collection with about 300 movies. I've been through so many rebuilds of its database over the years that I now have all my folders named nice to line up with the movie db and I just let it scrape for a bit and everything comes back without a bunch of manual cleanup anymore. As for plex I only had just installed it a couple days ago and hadn't ever really used it when I saw the constant disk spinning and went and found this thread. So nothing to lose there either.
April 6, 201412 yr Just for a quick short update. I reformatted and remounted without the discard option and it has been stable for three days running just my xbmc mysql database off of it. I'll try and reinstall plex today and verify it's ok I suspect it will be. I did verify that the drive is running a late enough firmware to support trim. But really the original vertex was kind of terrible yet the only SSD I could afford back in the day so this issue doesn't surprise me. I'll replace this thing eventually.
April 13, 201412 yr Thanks a lot for this guide I moved my ubuntu vm to an SSD using this method and it is flying. What a difference it made One quick question, though Normally they say that SSDs do a lot better when they are not full Is that the same when it comes to vm img files? My ubuntu vm is about 70GB. It is on a 120GB SSD. If I move my windows vm that is 40GB onto the same SSD, will it slow down significantly?? I heard that it is due to the SSD controller having a hard time finding spots to write new files on. But I am not sure how writes to img files are handled. Since there will only be two files on the SSD and no new files will be created??? I don't know enough to even guess
April 13, 201412 yr Author Generally yes, regardless of how it is being use (i.e. VM) ... but it also varies by SSD based on how much over-provisioning they do. That is, if they reserve some portion of the actual capacity outside what is accessible to the user. The more over-provisioned the more cells are available to recover from bad ones, and the more available for things like dealing with garbage collection. I can't give you a good rule of thumb, but yes I suppose I'd say much more than about 80% full and I'd start looking for a bigger SSD. Right, am image file is nothing special to the SSD. All it cares about is that some data needs to be written to some cells. If there are free empty cells great, if not then it has to erase an entire block first, the rewrite it with the data that was already there plus your new data.
April 13, 201412 yr Add a "force user = root" to the smb-extra.conf. This will allow writing to it. There might be other ways but this suggestion to me by RockDawg is the only one that I've used - it works. Thank You. Instructions updated. With force user = root, I was not able to access the share over the network. Windows prompted me for a username/password (root password didn't work) when I tried to access the share. I removed this line from the file, and I can access and read/write the disk via SMB just fine. Note I'm using 6.0-beta4.
April 18, 201412 yr First I would like to thank jumperalex for his remarkable tutorial. At the end of paragraph 1c, the process is validated by the following command: root@Tower:/dev/disk/by-id# fsck -N /dev/hd* fsck from util-linux-ng 2.17.2 [/sbin/fsck.ext2 (1) -- /dev/hde] fsck.ext2 /dev/hde [/sbin/fsck.reiserfs (2) -- /dev/hde1] fsck.reiserfs /dev/hde1 [/sbin/fsck.ext2 (3) -- /dev/hdf] fsck.ext2 /dev/hdf [/sbin/fsck.reiserfs (4) -- /dev/hdf1] fsck.reiserfs /dev/hdf1 [/sbin/fsck.ext2 (5) -- /dev/hdg] fsck.ext2 /dev/hdg [/sbin/fsck.reiserfs (6) -- /dev/hdg1] fsck.reiserfs /dev/hdg1 [/sbin/fsck.ext4 (7) -- /dev/hdh] fsck.ext4 /dev/hdh [/sbin/fsck.ext4 ( -- /dev/hdh1] fsck.ext4 /dev/hdh1 Everything looks normal so far, the SSD being /dev/hdh. But the next step shows the SSD device (first line), but not -part1 [quote]root@Tower:/dev/disk/by-id# ls ata-SSD2SC120G3LC709B121-460I_PNY14130000471460570@ ata-TOSHIBA_DT01ACA200_13E7689AS@ ata-TOSHIBA_DT01ACA200_13E7689AS-part1@ ata-TOSHIBA_DT01ACA200_13EDSNWGS@ ata-TOSHIBA_DT01ACA200_13EDSNWGS-part1@ ata-TOSHIBA_DT01ACA200_23JMGDZAS@ ata-TOSHIBA_DT01ACA200_23JMGDZAS-part1@ ata-TOSHIBA_DT01ACA200_Y2P4X60AS@ ata-TOSHIBA_DT01ACA200_Y2P4X60AS-part1@ ata-TOSHIBA_DT01ACA300_23DES3ZGS@ ata-TOSHIBA_DT01ACA300_23DES55GS@ ata-WDC_WD10EADS-11M2B1_WD-WCAV54129425@ scsi-SATA_TOSHIBA_DT01ACA_13E7689AS@ scsi-SATA_TOSHIBA_DT01ACA_13E7689AS-part1@ scsi-SATA_TOSHIBA_DT01ACA_13EDSNWGS@ scsi-SATA_TOSHIBA_DT01ACA_13EDSNWGS-part1@ scsi-SATA_TOSHIBA_DT01ACA_23JMGDZAS@ scsi-SATA_TOSHIBA_DT01ACA_23JMGDZAS-part1@ scsi-SATA_TOSHIBA_DT01ACA_Y2P4X60AS@ scsi-SATA_TOSHIBA_DT01ACA_Y2P4X60AS-part1@ usb-Corsair_Flash_Voyager_4cb768f2227108-0:0@ usb-Corsair_Flash_Voyager_4cb768f2227108-0:0-part1@ wwn-0x5000039ff3c23b52@ wwn-0x5000039ff3c23b52-part1@ wwn-0x5000039ff3c34620@ wwn-0x5000039ff3c34620-part1@ wwn-0x5000039ff3c5cd52@ wwn-0x5000039ff3c5cd52-part1@ wwn-0x5000039ff3c8d92d@ wwn-0x5000039ff3c8d92d-part1@ Of course, I tried the rest of the tutorial, but the mount cmd shows that the ext4 disk is not present root@Tower:/dev/disk/by-id# mount proc on /proc type proc (rw) sysfs on /sys type sysfs (rw) tmpfs on /var/log type tmpfs (rw,size=128m) /dev/sda1 on /boot type vfat (rw,noatime,nodiratime,umask=0,shortname=mixed) /dev/md1 on /mnt/disk1 type reiserfs (rw,noatime,nodiratime,user_xattr,acl) /dev/md2 on /mnt/disk2 type reiserfs (rw,noatime,nodiratime,user_xattr,acl) /dev/md3 on /mnt/disk3 type reiserfs (rw,noatime,nodiratime,user_xattr,acl) /dev/md4 on /mnt/disk4 type reiserfs (rw,noatime,nodiratime,user_xattr,acl) /dev/md5 on /mnt/disk5 type reiserfs (rw,noatime,nodiratime,user_xattr,acl) /dev/hdf1 on /mnt/cache type reiserfs (rw,noatime,nodiratime,user_xattr,acl) shfs on /mnt/user0 type fuse.shfs (rw,nosuid,nodev,noatime,allow_other) shfs on /mnt/user type fuse.shfs (rw,nosuid,nodev,noatime,allow_other) Where did I go wrong? ls_by-id.txt
April 18, 201412 yr Author Well shit ... um ... are you sure you did 1b successfully? i have ot admit despite writing the guide, I'm not a linux guru. Maybe someone else has seen this? What version of unraid are you using? I sure hope it isn't something with the new kernel on 6.0. I'm sure it would be fixable, just annoying. I wrote this based on what was in 5.x and probably 5.0.1 iirc
April 18, 201412 yr First, thank you for not treating me as a moron I am running a very standard UnRAID v5.0.5 on a decent hardware CPU AMD A8-5500 APU on MB Biostar A85W with 8 GB RAM (I know it is not relevant but just to say I am doing it by the book). I am trying to install a brand new PNY 120 GB SSD XR8. I haven't done any firmware modification (yet). Here's what I get after a reboot. The SSD is recognized, even partially /dev/hdh1 Array Disk Status Status Disk Mounted Device Model/Serial Temp Reads Writes Errors Size Used %Used Free OK parity /dev/hdg TOSHIBA_DT01ACA300_23DES3ZGS 39°C 87 32 OK /dev/md1 /mnt/disk1 /dev/sdc TOSHIBA_DT01ACA200_13EDSNWGS 35°C 749 10 2.00T 1.45T 73% 553.00G OK /dev/md2 /mnt/disk2 /dev/hde TOSHIBA_DT01ACA300_23DES55GS 37°C 6449 10 3.00T 2.25T 76% 746.17G OK /dev/md3 /mnt/disk3 /dev/sde TOSHIBA_DT01ACA200_13E7689AS 39°C 856 10 2.00T 1.26T 64% 736.15G OK /dev/md4 /mnt/disk4 /dev/sdd TOSHIBA_DT01ACA200_23JMGDZAS 38°C 1246 9 2.00T 1.20T 60% 803.19G OK /dev/md5 /mnt/disk5 /dev/sdb TOSHIBA_DT01ACA200_Y2P4X60AS 37°C 984 10 2.00T 1.39T 70% 613.34G Total: 11.00T 7.55T 68% 3.45T UNRAID Drive Device Model/Serial Mounted File System Size Used %Used Free /dev/sda1 usb-Corsair_Flash_Voyager_4cb768f2227108-0:0 /boot 8.07G 235.23M 3% 7.83G Cache Drive Device Model/Serial Mounted File System Temp Size Used %Used Free /dev/hdf1 /mnt/cache 34°C 1.00T 8.26G 1% 991.92G Drive Partitions - Not In Protected Array Device Model/Serial Mounted File System Temp Size Used %Used Free /dev/hdf WDC_WD10EADS-11M2B1_WD-WCAV54129425 34°C 1.00T /dev/hdh SSD2SC120G3LC709B121-460I_PNY14130000471460570 30064771072°C 120.03G /dev/hdh1 30064771072°C I must have screwed up somewhere, but frankly I don't see where. I have started over the whole section 1, including erasing the partition before creating a new one. Same result.
April 18, 201412 yr PROBLEM SOLVED !!!! ;D Actually, when putting back the motherboard to default safe values, it changed the SATA type and SATA IDE values. The correct values must be SATA TYPE ---> AHCI SATA IDE combined mode ---> DISABLED All of a sudden the drives hdX became sdX. And everything went back to normal as described in this one-of-its-kind tutorial. Again, thank you Jumperalex for such a great piece of work!
April 18, 201412 yr Author Excellent, I'm glad you figured it out. I'll have to file that one away in my troubleshooting Rolodex Sent from my Nexus 7 using Tapatalk
April 23, 201412 yr Would this work to add a nonssd outside of the array, and if not could you point me in the correct direction? Thanks
April 23, 201412 yr Would this work to add a nonssd outside of the array, and if not could you point me in the correct direction? Thanks Yes you could use this general guide to mount a drive outside the array. This was written specifically for SSD's to format ext4 allowing for TRIM utilization. Direct quote from OP: If you found this thread looking for more generic into on how to mount a drive outside the array it should still serve your pupose. There is no reason not to format as ext4 unless you have a specific reason not to. However, if you are NOT using an SSD then I would suggest you skip adding discard (aka trim) support. I don't know if it would actually hurt anything, but it is useless non the less and I don't like doing useless stuff Otherwise the rest of this thread should certainly still be useful and instructive.
April 23, 201412 yr Author exact same instructions with one option and one subtraction: Subtraction: Do NOT use DISCARD when mounting the drive, that is only for SSDs Option: you can now use just about any filesystem you want. The main reason for ext4, beyond being generally good for its purpose, is that it supports DISCARD. But there is no reason not to use ext4, like I said you now have options
April 26, 201412 yr Not sure what the issue was originally but it doesn't appear to be the drive mount script. However is there any way to view the drive from windows explorer? I'm not 100% sure if I need to create the directories for NZBGet prior to assigning them and wanted to see the folder structure in the explorer for the appdisk. For example I'm trying to create a NZBGet folder for NZBDrone to drop files into: /mnt/appdisk/nzbget/tv Getting a path does not exist error, not sure if I could view the appdisk in the windows explorer to create these folders?
May 7, 201412 yr Trying to go through the guide and format an ssd but I'm running into a problem right off the bat. I typed fdisk -cu /dev/sdb sdb is my 500GB ssd. But I get the following, root@Tower:~# fdisk -cu /dev/sdb Usage: fdisk [options] <disk> change partition table fdisk [options] -l <disk> list partition table(s) fdisk -s <partition> give partition size(s) in blocks Options: -b <size> sector size (512, 1024, 2048 or 4096) -c[=<mode>] compatible mode: 'dos' or 'nondos' (default) -h print this help text -u[=<unit>] display units: 'cylinders' or 'sectors' (default) -v print program version -C <number> specify the number of cylinders -H <number> specify the number of heads -S <number> specify the number of sectors per track root@Tower:~# which do not match the options you list in the guide. Checked to see what version of fdisk I'm running, it's root@Tower:~# fdisk -v fdisk (util-linux 2.21.2) I also checked to see if the brand new out of the box ssd is formatted and it returned root@Tower:~# fsck -N /dev/sdb1 fsck from util-linux 2.21.2 [/sbin/fsck.ext2 (1) -- /dev/sdb1] fsck.ext2 /dev/sdb1 How should I proceed?
May 7, 201412 yr Nevermind...I was able to continue by just doing fdisk /dev/sdb Then I got the same options as you had in the guide.
Archived
This topic is now archived and is closed to further replies.