|
ShaneB posted:All the file paths have changed to new linuxy names, but that's a more straightforward find/replace before trying to use it with the linux build. Is there any reason to choose rTorrent or Transmission over uTorrent? I'd like a really good web interface, as I'd like to run it headless if possible. Really good meaning: allow me to set destination directory when I add a torrent, something uTorrent didn't do last time I used the web interface. I can't remember if Transmission will let you do that but Transmission (and a few others like deluge) will let you connect your native client to another client/headless instance so you get the full interface without having to deal with an iffy ajax reimplementation of the client. For transmission this is only possible with the QT or 3rd party interface unless they've revamped the GTK interface since last time I used it. In deluge the default interface is able to do this.
|
#
?
Jan 4, 2012 22:26
|
|
|
|
| # ? May 1, 2024 23:36 |
|
|
Last time I tried the linux version of uTorrent it was kind of wonky and didn't work very well. I've switched to deluge and never looked back since. I use the web interface exclusively since my server is headless.
|
|
#
?
Jan 5, 2012 00:54
|
|
|
Is there a torrent client that will move the files once I remove them from the client. I've got it working how I want with Azureus, but running that "headless" is a pain in the rear end. What I want is: 1) Put .torrent file in watch directory 2) torrent starts, puts data in Incomplete folder (if it slurps the torrent file somewhere else that would be cool too) 3) Once torrent is done downloading, move the data into a Seeding folder (though this isn't all that necessary) 4) Once it's done seeding to a ratio I specify, stop the torrent. 5) Once I remove the torrent from the client, the data gets moved to a Completed folder. Rtorrent seems the one most likely to do this, but gently caress if I can figure it out without just reading the source code.
|
|
#
?
Jan 5, 2012 05:41
|
|
|
FISHMANPET posted:Is there a torrent client that will move the files once I remove them from the client. I've got it working how I want with Azureus, but running that "headless" is a pain in the rear end. I know deluge and transmission support moving torrents once they're done downloading. If you want to move a 2nd time after removal then that'll be more complicated but I think it's doable either with a plugin in deluge or via an 'execute script' command in either deluge or transmission.
|
|
#
?
Jan 5, 2012 06:19
|
|
|
ShaneB posted:Finally, finally, switched my super-crashy Windows 7 server (never got the crashing diagnosed, ever) over to Ubuntu 11.10 last night. Went to bed at 3:15am after getting most of the services I wanted up and running right. I have a few things left, such as a FTP server and a torrent client. Can't you just run µTorrent in Wine? Fresh µTorrent install, dirs configured to taste corresponding to new filetree, transfer filedata to new dirs, import old resume.dat, verify data.
|
|
#
?
Jan 5, 2012 11:02
|
|
|
ShaneB posted:All the file paths have changed to new linuxy names, but that's a more straightforward find/replace before trying to use it with the linux build. Is there any reason to choose rTorrent or Transmission over uTorrent? I'd like a really good web interface, as I'd like to run it headless if possible. Really good meaning: allow me to set destination directory when I add a torrent, something uTorrent didn't do last time I used the web interface. I'm a fan of Deluge (which is a headless daemon with web, command line, and GTK interfaces), but to be honest I've never used the web interface. The GTK interface lets you specify a destination directory, though.
|
|
#
?
Jan 5, 2012 15:49
|
|
|
Deluge is the bomb. I run the daemon on my box with all the storage, and then I connect to it from my laptop using the GTK client. By default deluge gtk app connects to localhost, but you just have to switch an option something like "classic mode", and then it will prompt you which host to connect to when it starts. I never bothered with the web interface because the GTK client does everything, and I prefer the native UI.
|
|
#
?
Jan 5, 2012 19:13
|
|
|
I have a linux programming question, I am not sure if this is the right thread. I'm trying to teach myself bash programming, using ubuntu 11.10. I've followed some website or another (forget the name). I made a script, gave it propor permissions via chmod, set up my /bin folder as a path. At first I was able to execute "my_script" via command line, and then I ran into a hung-up terminal from launching gedit from same command line. Now, I'm not sure how to kill whatever process (in this case gedit I suppose?) without just clicking "x" on the terminal. Things have not gone well. now I cannot run any of my bash scripts from my terminal. It just says "command not found". I tried running it as ./my_script in its folder with same results. 1) how do I locate hung-up process and kill it via terminal commands? 2) did I break bash by incorrectly ending a process? update: I'm an idiot. I rebooted and was fine. For sure I rather should have thought of that as solution #1. Oh well - for those that have my woes.. reboot, my friends. Reboot. DrWasabi fucked around with this message at 02:55 on Jan 6, 2012 |
|
#
?
Jan 6, 2012 01:40
|
|
|
DrWasabi posted:I have a linux programming question, I am not sure if this is the right thread. It sounds like you may have wiped the old PATH and completely replaced it with a new one instead of appending it. `echo $PATH` will give you all the paths to binaries, separated by a ":". If you want to add another directory, you need to append it to that line.
|
|
#
?
Jan 6, 2012 04:50
|
|
|
DrWasabi posted:update: I'm an idiot. I rebooted and was fine. For sure I rather should have thought of that as solution #1. Oh well - for those that have my woes.. reboot, my friends. Reboot. This is incorrect. Linux has very few problems that need a reboot. I suggest learning about the basics how linux (and most other unices) handle processes, shells and variables. What you should have done, is start another shell, use "ps -ef" (or "ps aux", depending on preference) to find out what the PID of your stuck process was, then kill it using "kill <pid>" How do you find what process to kill? "ps -ef | grep <script name>" (Before anyone starts about pkill, this way is better for learning how pids and stuff work) Furthermore, you terminal was "hung up" because you ran gedit in the foreground. Next time this happens, hit control-Z, then type "bg". And next time you run gedit fromt the terminal, run "gedit &" the & will put the process in the background, allowing you to keep using your terminal. spankmeister fucked around with this message at 09:45 on Jan 6, 2012 |
|
#
?
Jan 6, 2012 09:42
|
|
|
Is there an easy way to access a Windows share in Linux? I don't know if share is the right word actually. Basically I have a Windows box and I want it to share files with my Linux box but I'm not sure how to go about this.
|
|
#
?
Jan 6, 2012 11:38
|
|
|
Social Animal posted:Is there an easy way to access a Windows share in Linux? I don't know if share is the right word actually. Basically I have a Windows box and I want it to share files with my Linux box but I'm not sure how to go about this. Yes, Linux can access SMB shares. What distro do you use? Almost every modern desktop disto can access windows shares out of the box.
|
|
#
?
Jan 6, 2012 11:41
|
|
|
spankmeister posted:Yes, Linux can access SMB shares. What distro do you use? Almost every modern desktop disto can access windows shares out of the box. I'm currently using Fedora 16 (gnome3 fallback). I'm just not sure how to go about doing it as I'm pretty new to Linux in general. In the file manager I can see a network category with a browse network option. It shows up with Windows Network but it says "failed to retrieve share list from server." Is that a problem on the Windows box?
|
|
#
?
Jan 6, 2012 11:49
|
|
|
Social Animal posted:I'm currently using Fedora 16 (gnome3 fallback). I'm just not sure how to go about doing it as I'm pretty new to Linux in general. In the file manager I can see a network category with a browse network option. It shows up with Windows Network but it says "failed to retrieve share list from server." Is that a problem on the Windows box? Have you shared any folders on the windows box?
|
|
#
?
Jan 6, 2012 12:07
|
|
|
Social Animal posted:I'm currently using Fedora 16 (gnome3 fallback). I'm just not sure how to go about doing it as I'm pretty new to Linux in general. In the file manager I can see a network category with a browse network option. It shows up with Windows Network but it says "failed to retrieve share list from server." Is that a problem on the Windows box? Older versions of Windows used to give a list of their shared folders to anyone who asked on the network, but I think the newer versions might be more like "identify yourself first  ". I personally prefer KDE, so I cannot tell the exact location, but somewhere in the GNOME settings there might/should be a way to configure a Windows username and password which will be presented to the Windows servers when looking for network shares. I expect at least the password field will be empty by default. (On this Debian+KDE system, it's slightly illogically at System Settings -> Sharing, which you might expect be related to sharing stuff to others, not accessing other systems' shares. Meh.) Specify a Windows username that can be expected to make sense to the Windows box (i.e. you may need to use the form DOMAIN\username in some cases) and the password associated with it, then try again.
|
|
#
?
Jan 6, 2012 13:42
|
|
|
ToxicFrog posted:I'm a fan of Deluge (which is a headless daemon with web, command line, and GTK interfaces), but to be honest I've never used the web interface. The GTK interface lets you specify a destination directory, though. Deluge was perfect, thanks for the suggestion. I just manually re-started the torrents and specified the locations, but the daemon/headless setup is choice.
|
|
#
?
Jan 6, 2012 20:59
|
|
|
When multiple processes are writing files to a single disk, are the writes interleaved in a file basis or sector basis? They way I figure, a single hard drive or even an array can only write one stream at a time so if a file is ready to be written in its entirety, it would be more efficient to write that file completely and then write the next file from the same or another process instead of writing a peice of a file1 and then a peice of file2 and then back to another peice of file1. I guess this all gets complicated when the process isn't ready to write the whole file but opens the file for writing.
|
|
#
?
Jan 6, 2012 22:58
|
|
|
Shaocaholica posted:When multiple processes are writing files to a single disk, are the writes interleaved in a file basis or sector basis? They way I figure, a single hard drive or even an array can only write one stream at a time so if a file is ready to be written in its entirety, it would be more efficient to write that file completely and then write the next file from the same or another process instead of writing a peice of a file1 and then a peice of file2 and then back to another peice of file1. I guess this all gets complicated when the process isn't ready to write the whole file but opens the file for writing. Writes are sent to the kernel as sequences of filesystem operations. The kernel translates them into sector operatons, and reorders them such that they can be performed as quickly as possible. So, files are written to disc sector-at-a-time, a given sector may have data for more than one file depending on the filesystem involved, most filesystems do not requires that files be stored contiguously on disc, and "later" writes may be committed to disc before "earlier" writes are because the kernel finds it faster that way. This is assuming none of the processes involved invoke any kind of barrier or direct mapping, which introduces a whole new level of complication in IO scheduling.
|
|
#
?
Jan 6, 2012 23:07
|
|
|
ShoulderDaemon posted:Writes are sent to the kernel as sequences of filesystem operations. The kernel translates them into sector operatons, and reorders them such that they can be performed as quickly as possible. So, files are written to disc sector-at-a-time, a given sector may have data for more than one file depending on the filesystem involved, most filesystems do not requires that files be stored contiguously on disc, and "later" writes may be committed to disc before "earlier" writes are because the kernel finds it faster that way. Does the kernel try to keep files contiguous when possible? Is that something thats even taken into consideration in the IO scheduler?
|
|
#
?
Jan 6, 2012 23:12
|
|
|
IO scheduling is very complex, and Linux has a few different schedulers that all have their own advantages and disadvantages. None are as dumb as to only write one file at a time though. 
|
|
#
?
Jan 6, 2012 23:12
|
|
|
Shaocaholica posted:Does the kernel try to keep files contiguous when possible? Is that something thats even taken into consideration in the IO scheduler? That's a filesystem issue not an IO scheduler issue.
|
|
#
?
Jan 6, 2012 23:13
|
|
|
spankmeister posted:None are as dumb as to only write one file at a time though.
|
|
#
?
Jan 6, 2012 23:16
|
|
|
Shaocaholica posted:Well its not so much dumb but there are certain scenarios where this would be appropriate. True indeed, but in most cases it's not and I meant that the kernel has some intelligent algorithms to determine how to write something to the disk instead of "write file 1, write file 2, write file 3" etc...
|
|
#
?
Jan 6, 2012 23:21
|
|
|
The IO schedule chooses which order to send the read or write commands to the disk. The actual location of the sectors read or written is decided by the file system driver (or LVM/device-mapper, the userspace app directly accessing the block device, etc.) Filesystems are designed to minimize fragmentation if possible.
|
|
#
?
Jan 6, 2012 23:22
|
|
|
pseudorandom name posted:Some Filesystems are designed to minimize fragmentation if possible. Fixed that for you. I doubt any log based filesystem is going to do anything about fragmentation. With the move to SSDs and such, fragmentation is something you can ignore to improve other areas of a filesystem.
|
|
#
?
Jan 6, 2012 23:24
|
|
|
pseudorandom name posted:Filesystems are designed to minimize fragmentation if possible. Well there's also the overhead of switching from one file to another before its finished which can lead to slower aggregate writes. I'm sure smarter people have come up with logic for this but I justed wanted to ask to see if anyone could explain that logic.
|
|
#
?
Jan 6, 2012 23:30
|
|
|
The order applications submit requests to the kernel isn't the order they're submitted to the drive. That's the whole point of the IO scheduler, it reorders requests (where possible) to minimize seeks.
|
|
#
?
Jan 6, 2012 23:42
|
|
|
Ok, here's a simple case. Copying 10 dirs from one physical drive to another using cp. However, 10 instances of cp are lauched at the same time and allowed to run concurrently. All operations create new files so we don't have to worry about old file locations/sizes. What kind of bahavior can we expect from the IO scheduler given EXT3/EXT4 destination file system?
|
|
#
?
Jan 6, 2012 23:45
|
|
|
Shaocaholica posted:Well there's also the overhead of switching from one file to another before its finished which can lead to slower aggregate writes. I'm sure smarter people have come up with logic for this but I justed wanted to ask to see if anyone could explain that logic. This comment makes a lot of assumptions about the underlying filesystem and storage media. Log based filesystems, which I've previously mentioned, won't have this problem. Nor will writing on any media that doesn't seek. In the meantime you try to refresh your browser and whoops, can't write to disk so your whole UI just freezes up.
|
|
#
?
Jan 6, 2012 23:46
|
|
|
Shaocaholica posted:Ok, here's a simple case. Copying 10 dirs from one physical drive to another using cp. However, 10 instances of cp are lauched at the same time and allowed to run concurrently. All operations create new files so we don't have to worry about old file locations/sizes. You can expect the IO scheduler to sort the read commands to the source media and the write commands to the destination media (which are in two separate queues) to minimize seeks as much as possible, given the ordering constraints necessary to satisfy the ext3/ext4 integrity guarantees.
|
|
#
?
Jan 6, 2012 23:56
|
|
|
A little old but it describes what the IO scheduler is trying to do and how a couple work: http://www.linuxjournal.com/article/6931
|
|
#
?
Jan 7, 2012 00:58
|
|
|
taqueso posted:A little old but it describes what the IO scheduler is trying to do and how a couple work: http://www.linuxjournal.com/article/6931 Neato. Seems like you can set a different IO scheduler per disk: http://www.cyberciti.biz/faq/linux-change-io-scheduler-for-harddisk/
|
|
#
?
Jan 7, 2012 03:16
|
|
|
I recently bought a HP ENVY 14-2020NR and I'm trying to get ubuntu to work on the thing. I don't like the new ubuntu 11 so I'm sticking with 10.04 LTS. Anyway, it seems like there's a lot of problems regarding drivers. First of all, the graphics card is a Radeon HD 6630M, which I never got to work--ubuntu found some drivers which after installation caused ubuntu to no longer boot. (This was a problem with both ubuntu 11 and 10.04, I tried installing both.) Ubuntu uses the integrated card that comes with the motherboard, which is pretty lovely--the resolution is stuck at 1024x768 and I can't adjust the brightness, so I'd like to try to fix that. Since then I've reinstalled ubuntu and now there's no networking (which I at least had before). I've poked around on google and here are some things I've done: code:I gathered that the driver I need is this one: http://www.broadcom.com/support/802.11/linux_sta.php so I download it, then unzip, and in that folder I do code:code:ran code:
|
|
#
?
Jan 7, 2012 03:51
|
|
|
Install a newer ubuntu version with a newer kernel. If you don't like unity (the new UI) try out xubuntu or lubuntu.
|
|
#
?
Jan 7, 2012 04:11
|
|
|
Ugh, NFSv4. I've got a Solaris 11 Express server sharing some stuff with NFS. Because I used NFSv4 ACLs for CIFS sharing, the server started exporting NFSv4. On my old Ubuntu 10.04 box I somehow got mounting to work, but the PSU died and now I'm virtualizing and installed a fresh copy of Ubuntu 11.10, and now I don't know how to get mounting working. I was able to mount the entire tree (which I don't really like, or understand why they took away the ability to mount particular directories for NFSv4) but none of the UID/GIDs map. What the hell why is NFSv4 so lovely.
|
|
#
?
Jan 7, 2012 09:18
|
|
|
What commands should I run on a remote server to ensure that the next tenant doesn't try to get creative with data recovery tools?
fletcher fucked around with this message at 10:43 on Jan 7, 2012 |
|
#
?
Jan 7, 2012 10:36
|
|
|
telcoM posted:Older versions of Windows used to give a list of their shared folders to anyone who asked on the network, but I think the newer versions might be more like "identify yourself first
|
|
#
?
Jan 7, 2012 11:32
|
|
|
fletcher posted:What commands should I run on a remote server to ensure that the next tenant doesn't try to get creative with data recovery tools? I assume you have SSH access. Do you have console access and/or physical access to the host?
|
|
#
?
Jan 7, 2012 11:54
|
|
|
hackedaccount posted:I assume you have SSH access. Do you have console access and/or physical access to the host? The only access to the machine is SSH.
|
|
#
?
Jan 7, 2012 11:59
|
|
|
|
| # ? May 1, 2024 23:36 |
|
|
Social Animal posted:Is there an easy way to access a Windows share in Linux? I don't know if share is the right word actually. Basically I have a Windows box and I want it to share files with my Linux box but I'm not sure how to go about this. You can always mount your shares the classic way, as either manually from the terminal, or an appropriate entry in /etc/fstab. Typing from memory, a terminal entry could look like this: "mkdir /mnt/mywindowsshare && mount -t cifs \\windowshost\someshare /mnt/mywindowsshare -o user=windowsuser%passsword " Look it up with "man mount.cifs" or "man fstab". fletcher posted:What commands should I run on a remote server to ensure that the next tenant doesn't try to get creative with data recovery tools? You may want to start with clearing the local history ("history -c") in case you entered any sensitive information (cleartext user / password combos). Shaocaholica posted:Neato. Additional info I found thanks to this very thread: http://google-opensource.blogspot.com/2008/08/linux-disk-scheduler-benchmarking.html http://www.ertos.nicta.com.au/publications/#ug ("I/O scheduling on RAID")
|
|
#
?
Jan 7, 2012 12:17
|
|
