Mike, Netflix streams videos like YouTube. using a Wii, it will only use ~240p video quality which works fine with a 128K download rate. unfortunately, for downloading large files or games, no go unless you like waiting 8 hours to download 900MB or worse, a week to download DC Universe Online via PS3. (DCU is 15GB)
I'm currently attempting a 3G download of the ISO of Kubuntu as we speak on my Android phone. but given that it's only downloading at roughly 50Kbps, it is SLOWER than my home internet. my home internet is down, modem dies if it gets into the 20s in outside temp (modem is outdoor and attached to an antenna)
do I need to put my alias in a bash file to run at startup in order to make them work?
If you are using bash, edit your ~/.bashrc file and add one alias per line. Once you save and close the file, run this to load your new aliases immediately:
source ~/.bashrcOtherwise, the new aliases will load whenever you open a new instance of the shell.
*nix has never used 'extensions' as in the DOS/windows world. UNIX and its offspring actually determine a file's type based on its contents, NOT on its 'extension.'
The reason I keep putting 'extension' in quotes is because in *nix there really is no such thing as a file extension. Back in the DOS 8.3 (file names limited to 8 characters, plus a MANDATORY dot-plus-3-character extension), in UNIX we had no such ridiculous limitations.
I've never used a version of MS/PC-DOS that actually required any filename extension, much less a full 3 characters.
The filesystems for earlier versions of UNIX®, that were developed to run on early minicomputers with roughly the same computing resources as a modern day wristwatch, were in fact quite constrained. Let's not fall into a trap such as "UNIX was always as it is now".
The best-known UNIX® method of determining file type, the magic number, was introduced with Version 7, nearly a decade after UNIX/C was conceived. Obviously it wasn't the only one! Later on, as the Internet became more of a factor in computing, filename extensions became more important in determining remote file types. Today the most advanced POSIX desktop managers typically use filename extensions to associate applications with.
The Internet has traditionally been a vehicle for inclusion, not exclusion. Along those lines, Internet protocols like FTP and HTTP have been available to users of "lowly" operating systems like 16-bit DOS/WIN, and the Internet has even adapted some things like filename extensions to enhance system interoperability in its open environment. IMHO things like this should be celebrated, not derided.
Ubuntu isn't the download i worry about. i only have 128K down and maybe .5K up. it's slow. it was a rather long download on cable at my mom's house. my speed is roughly half that of DSL and 1/3 that of cable. i live in a trailer in the woods, it's a wireless system and i cannot get any faster speed.
One thing that you can do that will also help support the development of the FOSS that you enjoy is to order the physical discs to be delivered to you. Doing it this way provides a tiny bit of financial support to these people who often work for free on this stuff, and may end up getting you your software faster than possible by using a slow Internet connection. I used to do that before I got my first DSL circuit.
Was asked to associate a program with opening a file. I wanted FX to open it, but FX wasn't in the choice box. What do I use to associate with? In Windows it would be FX.ex- (left the last e off) What's operative in Ubuntu? Still running 11.04, will be changing to Mint, soon.
First, you must get out of your head the [idiotic] micro$oft way of doing things. Read some of my other comments in this thread about file 'extensions' and how they're basically meaningless in UNIX/Linux.
In *nix, it's a file's PERMISSIONS that indicate whether or not it's executable, not its extension. Therefore, a file with ANY name can be executable, such as ThisFile, as long as its permissions include executing it.
I don't know what FX is, but how do you normally fire it up? You said "Was asked to associate a program with opening a file. I wanted FX to open it..." Where were you when you were asked to associate a program with opening a file? I mean what program was asking you for that info?
I've never used a version of MS/PC-DOS that actually required any filename extension, much less a full 3 characters.
I understood that DOS couldn't *do* anything with a file unless it had an appropriate extension. For example, a 'batch' file couldn't be executed unless it had a .bat extension, a compiled program couldn't run unless it had .exe as its extension, a JPEG couldn't be displayed unless it had .jpg as its extension, etc. So you're saying this was NOT the case?
Quote:
The filesystems for earlier versions of UNIX®, that were developed to run on early minicomputers with roughly the same computing resources as a modern day wristwatch, were in fact quite constrained. Let's not fall into a trap such as "UNIX was always as it is now".
I guess I can clarify from now on that I'm speaking from 1985 on--when *I* started using *nix.
Quote:
The Internet has traditionally been a vehicle for inclusion, not exclusion. Along those lines, Internet protocols like FTP and HTTP have been available to users of "lowly" operating systems like 16-bit DOS/WIN, and the Internet has even adapted some things like filename extensions to enhance system interoperability in its open environment. IMHO things like this should be celebrated, not derided.
Were you using Usenet when windows users first showed up? I was. With their non-standards-compliant, so-called newsgroup clients--that did things like put background images in their posts...or attempted to, anyway...those of us using real newsgroup clients just saw a bunch of m$ garbage.
M$ has never believed in inclusion, hence its ongoing failure to comply with accepted standards for things like browsers, its proprietary plug-ins, such as Silverlight (which keeps me from using Netflix), the federal lawsuits brought against it for its uncouth, often illegal, business practices, etc. IMNSHO, that should be derided, not celebrated.
Device(s): HTC EVO 4g(Retired), HTC EVO LTE(Retired), LG Lucid(Returned), Motorola Droid Razr M, Nexus 7
Carrier: GOD
Thanks: 3,752
Thanked 3,962 Times in 1,980 Posts
lets say i want to move a large amount of folders and files over from one drive to another drive. besides the obvious drag and drop what would be the best way to do this in CLI? would it be to use the mv command and if so what would be the recommended way?
Hi, another question. when using my Toshiba (ubuntu plus kde) it seems to be a pain to boot up. today i put it in place my the Acer (it died, so i tossed it--was on its last legs anyway--keyboard didn't even work!) and first boot it locked up. second boot it got to KDE then froze. third boot it got stuck in an fsck loop (scanned, fixed errors, reboot, repeat cycle) and the fourth time it froze again, fifth time it finally got in, and worked fine. still works fine. this glitch happened last time too, the fsck loop at least. not sure what is going on. it claims the file system was unclean even though i did a proper shut down prior. also, if i let it try booting with the Kubuntu logo, it freezes during the boot process. i have to hit the 'esc' key to show verbose mode before it will work.
never runs out of RAM, perhaps i should just get into the habit of hibernating it instead of shutting it down?
Device(s): HTC EVO 4g(Retired), HTC EVO LTE(Retired), LG Lucid(Returned), Motorola Droid Razr M, Nexus 7
Carrier: GOD
Thanks: 3,752
Thanked 3,962 Times in 1,980 Posts
Quote:
Originally Posted by nickdalzell
Hi, another question. when using my Toshiba (ubuntu plus kde) it seems to be a pain to boot up. today i put it in place my the Acer (it died, so i tossed it--was on its last legs anyway--keyboard didn't even work!) and first boot it locked up. second boot it got to KDE then froze. third boot it got stuck in an fsck loop (scanned, fixed errors, reboot, repeat cycle) and the fourth time it froze again, fifth time it finally got in, and worked fine. still works fine. this glitch happened last time too, the fsck loop at least. not sure what is going on. it claims the file system was unclean even though i did a proper shut down prior. also, if i let it try booting with the Kubuntu logo, it freezes during the boot process. i have to hit the 'esc' key to show verbose mode before it will work.
never runs out of RAM, perhaps i should just get into the habit of hibernating it instead of shutting it down?
sure your not having a hard drive issue? that is what it is sounding like to me.
it is a brand new laptop. it had Win8 previously. the Acer's hard disk got into an activity loop and would not scan and i got the constant access light. that's not the problem here, it is like it has a process crash during load. Ubuntu with Unity never did this it started soon as KDE got installed. perhaps driver issue? it works fine if i never shut it down, or if i only hibernate or suspend. it's specifically when i restart or turn it off and back on (old habit)
ALSO, if i hit 'C' to skip the fsck it most of the time boots up fine-but i have to be in verbose mode anyway.
also got this weird prompt asking me for a 'kdewallet' password even though i never set up any 'wallets' or 'e-wallets' in KDE
lets say i want to move a large amount of folders and files over from one drive to another drive. besides the obvious drag and drop what would be the best way to do this in CLI? would it be to use the mv command and if so what would be the recommended way?
I'd suggest first looking at the options for mv:
Code:
man mv
Once you see what your choices are, you can decide what you want to do. For example, do you want it to work interactively? Do you want to see the files' names as they're being moved? Etc.
Read its man page and then post again if you need help figuring out how to do what you want.
I understood that DOS couldn't *do* anything with a file unless it had an appropriate extension. For example, a 'batch' file couldn't be executed unless it had a .bat extension, a compiled program couldn't run unless it had .exe as its extension, a JPEG couldn't be displayed unless it had .jpg as its extension, etc. So you're saying this was NOT the case?
For the sake of clarity and continuity, let's go back to what I was replying to:
Quote:
Originally Posted by MoodyBlues
*nix has never used 'extensions' as in the DOS/windows world. UNIX and its offspring actually determine a file's type based on its contents, NOT on its 'extension.'
The reason I keep putting 'extension' in quotes is because in *nix there really is no such thing as a file extension. Back in the DOS 8.3 (file names limited to 8 characters, plus a MANDATORY dot-plus-3-character extension), in UNIX we had no such ridiculous limitations. If I wanted to name something my.file.for.work, so be it. The dots were, and are, nothing but characters, like any other standard character. They didn't, and don't, dictate a file's type. *nix basically evolved over the years to accommodate windows users who thought of files as having extensions to dictate what they were.
Don't believe me? Do this: Let's say you have a JPEG file that's named your_pretty_sunset.jpg. Rename it simply your_pretty_sunset. Then fire up your preferred file manager (I use Dolphin), and watch it display the file's preview. Do the same on a windows box and what will happen? Nothing. Without the .jpg extension, it won't recognize it as a JPEG file.
Note that the post "shouted" things like "MANDATORY" that simply aren't true. I was just providing the correct information.
Yes, it's possible for an application program (regardless of the platform it runs on) to read magic numbers and use them to guess file types when there is no fillename extension. But that's not an OS feature; it's an application feature.
In my own particular situation, using the K Desktop Environment, it is in fact the filename extension that determines what application opens it.
When discussing these matters, it would be helpful to remember the role of the filesystem, operating system (and even desktop environment) and application program, and understand the capabilities and limitations of each in the greater system. The post that I addressed lacks that distinction.
I personally don't see any good coming from constantly bashing a rival, especially when it involves falsehood. I guess this forum doesn't forbid it, but maybe it would be more appropriate to start a "Microsoft Haters" topic rather than muddy this "Linux questions..." thread with stuff that's not being helpful to the topic at hand, or for Windows users who want to learn about Linux. That's just my opinion; take it or leave it.
if i wanted to wait for a few days to have discs delivered then i would not be concerned with an 8 hour download...
I hope I'm not putting words in Speed's mouth, but I think the point being made is that by ordering discs you're helping [in a small way] to financially support the software you're using. And, of course, getting discs so you don't have to bother with a long download.
lets say i want to move a large amount of folders and files over from one drive to another drive. besides the obvious drag and drop what would be the best way to do this in CLI? would it be to use the mv command and if so what would be the recommended way?
One practice that I often preach but don't practice myself is to keep the source file(s) until you're sure that they all made it to their destination intact. So instead of mv, you might do a cp, check to see that the files are there, and then rm the source files. This takes longer and requires more disk space (which may be a problem if your reason to move the files is a full disk). But it's the safest way to do it.
For the sake of clarity and continuity, let's go back to what I was replying to:
<< snip >>
Speed, I respect you too much to want to get into anything ugly.
So I'm going to wrap this up with this: If you do what I suggested, i.e., use the file command on a file with *NO* extension, you'll see that *nix accurately analyzes the type of file; I do not believe that any m$ OS has ever been able to do that. I also believe that m$ OSes cannot do anything with a file if it doesn't have an extension telling it what to do. So by "mandatory" I guess I really meant that an extension was mandatory *IF* you wanted to do anything with the file, like execute it.
I hope I'm not putting words in Speed's mouth, but I think the point being made is that by ordering discs you're helping [in a small way] to financially support the software you're using. And, of course, getting discs so you don't have to bother with a long download.
Yes, that and the fact that spending many hours trying to download a large ISO image, only to find that it doesn't work is no fun at all. IME a single disc sent by First Class mail takes 2 days to arrive, which can be faster than several download attempts. And you get a nice silk screened disk to boot!
Speed, I respect you too much to want to get into anything ugly.
I feel the same way. Thanks.
Quote:
So I'm going to wrap this up with this: If you do what I suggested, i.e., use the file command on a file with *NO* extension, you'll see that *nix accurately analyzes the type of file; I do not believe that any m$ OS has ever been able to do that.
Did you know that you can get a `file' command for DOS and Windows? No it's not provided by default from Microsoft. But by the same token, the `file' utility isn't necessarily installed by default. And in the case of Linux, `file' is one of the many utilities supplied by GNU and other sources.
My point is and has been that magic numbers are properties of the files themselves. It's not something that the OS adds, like resource forks in MacOS. Any OS or application can use magic numbers if they want to. But in this Internet age we use filename extensions more than ever. And it works surprisingly well.
Quote:
I also believe that m$ OSes cannot do anything with a file if it doesn't have an extension telling it what to do. So by "mandatory" I guess I really meant that an extension was mandatory *IF* you wanted to do anything with the file, like execute it.
Yes, Microsoft operating systems require file extensions for a file to be executable, just like UNIX® and work-alikes require an execution bit to be set in the filesystem, and VMS requires a ".COM" extension to signify COMmands. I don't see any good coming from calling one "good" and another "bad" though.
Once again, confusing the OS with data files and the application programs that use those data files serves no constructive purpose.
figured out the pattern needed to avoid boot looping. if i wait for the Kubuntu load screen to show, immediately hit escape, then hit C, then I, it will skip the disk check and boot up just fine. something is wrong with fsck. not sure how to disable it. it keeps insisting the filesystem has errors even though nothing is wrong
if i let fsck continue and scan it will scan, say checking inodes, etc. then all of a sudden (can't see the text before it does) reboot, and do the same thing. says 'filesystem was uncleanly mounted, check forced' and just do it over and over. i used to just hit C to cancel, but doing that just freezes the system. if i hit C for cancel and I to ignore, it cancels the disk check and everything is fine. is there anything i can edit in /etc/fstab to get rid of fsck so i can just do it in a terminal?
Device(s): HTC EVO 4g(Retired), HTC EVO LTE(Retired), LG Lucid(Returned), Motorola Droid Razr M, Nexus 7
Carrier: GOD
Thanks: 3,752
Thanked 3,962 Times in 1,980 Posts
Quote:
Originally Posted by MoodyBlues
I'd suggest first looking at the options for mv:
Code:
man mv
Once you see what your choices are, you can decide what you want to do. For example, do you want it to work interactively? Do you want to see the files' names as they're being moved? Etc.
Read its man page and then post again if you need help figuring out how to do what you want.
Quote:
Originally Posted by Speed Daemon
One practice that I often preach but don't practice myself is to keep the source file(s) until you're sure that they all made it to their destination intact. So instead of mv, you might do a cp, check to see that the files are there, and then rm the source files. This takes longer and requires more disk space (which may be a problem if your reason to move the files is a full disk). But it's the safest way to do it.
going with daemon's post I looked at mv but wasnt sure would rsync be better than cp? or am i thinking to much?
Device(s): Three Chinese Androids: Kliton I806LS, Lenovo P700i, Ampe A76
Carrier: China Mobile, China Telecom.
Thanks: 1,376
Thanked 1,190 Times in 831 Posts
Quote:
Originally Posted by nickdalzell
if i wanted to wait for a few days to have discs delivered then i would not be concerned with an 8 hour download...
I'm not concerned about 8 hour downloads myself, set the download going, usually bittorrent, go to bed, sleep 8 hours, or go to work for the day, download is finished. Plus the idea of getting CDs or DVDs shipped from the US or EU to China, could be expensive and lengthy. Plus I don't have a CD or DVD drive anyway. Even if the download took two weeks, it would still be quicker than posting discs here.
__________________
The People's Guide to Android in the People's Republic.
Honorary Grand Poobah Shenzhen University English Corner. http://welcometomychina.tumblr.com/ There are nine million bicycles in Beijing.
There are nine million Androids in Shenzhen.
Last edited by mikedt; January 21st, 2013 at 05:53 PM.
i would rather hotspot hop until i find a free spot that is much faster (one is only four miles from me, another at Target, and so on) than waste eight hours and be unable to do anything else online during that time. and if the download fails....like it did on my Android phone (angry still as i was 98% done! now it just says 'download interrupted cannot resume')
The Following User Says Thank You to nickdalzell For This Useful Post:
going with daemon's post I looked at mv but wasnt sure would rsync be better than cp? or am i thinking to much?
Choices, choices.
Any of them will work. Personally, I would--and I do--use mv. I normally use it in its default mode, which is silent (i.e., no visual output), but sometimes I'll add its -v argument so I can see the files' names as they're being moved.
Device(s): HTC EVO 4g(Retired), HTC EVO LTE(Retired), LG Lucid(Returned), Motorola Droid Razr M, Nexus 7
Carrier: GOD
Thanks: 3,752
Thanked 3,962 Times in 1,980 Posts
Quote:
Originally Posted by MoodyBlues
Choices, choices.
Any of them will work. Personally, I would--and I do--use mv. I normally use it in its default mode, which is silent (i.e., no visual output), but sometimes I'll add its -v argument so I can see the files' names as they're being moved.
just an update on my fsck issue. all the lockups, boot loops and various other errors were related to my root filesystem being stuck in read-only mode. for some reason the 'auto' fsck check would lock the system up, hard booting needed which would just keep it mounted read-only. what i had to do was cancel auto-fsck and drop to a manual recovery shell, entered fsck -C /dev/sda1 and it fixed everything. my hard drive was not bad or faulty, just some 'inodes' out of place
The Following User Says Thank You to nickdalzell For This Useful Post:
Device(s): Three Chinese Androids: Kliton I806LS, Lenovo P700i, Ampe A76
Carrier: China Mobile, China Telecom.
Thanks: 1,376
Thanked 1,190 Times in 831 Posts
Quote:
Originally Posted by MoodyBlues
It's been a while, but I've seen that happen with widgets in the past, and I honestly can't recall actually SOLVING the problem--instead I just deleted them from my panel. *shrug*
I just tried Nick's nixie clock widget for KDE Plasma, didn't work for me either. "Script initialization failed.", and I do have the required Python dependencies installed. This is on KDE 4.8.5. deleted it....*shrug* LOL.
Actually i did not 'tinker' the problem happened when i installed Kubuntu Desktop. when i shut it down (it did perform a proper shutdown) and turned it back on, it got stuck at the Kubuntu screen (the graphic logo) and i force shut down (which is likely what caused the file system to be 'unclean' to begin with) and had been cancelling the auto-fsck because it would always freeze. eventually it got to a point that cancelling it was not enough, and i had to hit 'i' for 'ignore errors' (fsck would terminate and say filesystem still had errors) and it apparently only mounted root read-only. otherwise other than random bugs it would boot up, browse the web and so on. (the Compiz effects i posted in Show your Desktop were accomplished even though the FS was read-only) but i could not access terminal, or install anything. i had to find out myself how to dump it into a recovery shell (this used to happen automatically long ago???) and manually fsck it and fix it
i am not sure if it is the framebuffer or what, but IF i boot up and have the graphical Kubuntu loading animation the computer WILL freeze. i still need to hit escape for verbose mode
if i let fsck continue and scan it will scan, say checking inodes, etc. then all of a sudden (can't see the text before it does) reboot, and do the same thing. says 'filesystem was uncleanly mounted, check forced' and just do it over and over. i used to just hit C to cancel, but doing that just freezes the system. if i hit C for cancel and I to ignore, it cancels the disk check and everything is fine. is there anything i can edit in /etc/fstab to get rid of fsck so i can just do it in a terminal?
Strange.
Assuming that the filesystem has always been OK and no data has been lost or mangled, then the init script (or whatever replaces init in your distro) that checks for the dirty bit and runs fsck must be wrong. I suppose that you can edit it. I can't think of anything that would leave the dirty bit set on a clean shutdown or reboot.
Editing /etc/fstab will not help; you need to find the boot script that's malfunctioning. On my own (OpenSUSE) system, I did a `grep fsck /etc/rc.d/*' (all of my init scripts are in that directory) and saw several likely culprits:
Code:
me@mypc:~> grep fsck /etc/rc.d/*
/etc/rc.d/boot:if rc_cmdline forcefsck > /dev/null || test -e /forcefsck ; then
/etc/rc.d/boot.cleanup:# Required-Start: boot.rootfsck
/etc/rc.d/boot.cleanup: rm -f /tmp/.X*lock /var/spool/uucp/LCK* /var/log/sa/sadc.LOCK /fsck_corrected_errors
/etc/rc.d/boot.clock:# Required-Start: boot.rootfsck boot.udev
/etc/rc.d/boot.clock:# Required-Stop: boot.rootfsck
grep: /etc/rc.d/boot.d: Is a directory
/etc/rc.d/boot.device-mapper:# Required-Start: boot.udev boot.rootfsck
/etc/rc.d/boot.dmraid:# Required-Start: boot.udev boot.rootfsck
/etc/rc.d/boot.dmraid:# Required-Stop: boot.udev boot.rootfsck
/etc/rc.d/boot.klog:# Required-Start: boot.rootfsck
/etc/rc.d/boot.localfs:# Required-Start: boot.udev boot.rootfsck
/etc/rc.d/boot.localfs:# to get max number of parallel fsck processes
/etc/rc.d/boot.localfs: # do fsck and start sulogin, if it fails.
/etc/rc.d/boot.localfs: # skip fsck if running on battery
/etc/rc.d/boot.localfs: if test -f /forcefsck -o "$DO_FORCEFSCK" = "yes" ; then
/etc/rc.d/boot.localfs: # If we use a serial console, don't use the fsck progress bar
/etc/rc.d/boot.localfs: fsck $FSCK_OPTS
/etc/rc.d/boot.localfs: echo "fsck succeed, but reboot is required."
/etc/rc.d/boot.localfs: echo "fsck failed for at least one filesystem (not /)."
/etc/rc.d/boot.localfs: > /fsck_corrected_errors
/etc/rc.d/boot.localfs: rm -f /etc/nologin /nologin /fastboot /forcefsck /success
/etc/rc.d/boot.localnet:# Required-Start: boot.rootfsck
/etc/rc.d/boot.lvm:# Required-Start: boot.udev boot.rootfsck
/etc/rc.d/boot.md:# Required-Start: boot.udev boot.rootfsck
/etc/rc.d/boot.rootfsck:# /etc/init.d/boot.rootfsck
/etc/rc.d/boot.rootfsck:# Provides: boot.rootfsck
/etc/rc.d/boot.rootfsck:# to get max number of parallel fsck processes
/etc/rc.d/boot.rootfsck: # fsck may need a huge amount of memory, so make sure, it is there.
/etc/rc.d/boot.rootfsck: # do fsck and start sulogin, if it fails.
/etc/rc.d/boot.rootfsck: if test -f /forcefsck -o "$DO_FORCEFSCK" = "yes" ; then
/etc/rc.d/boot.rootfsck: # skip fsck if running on battery
/etc/rc.d/boot.rootfsck: # If we use a serial console, don't use the fsck progress bar
/etc/rc.d/boot.rootfsck: fsck $FSCK_OPTS $ROOTFS_BLKDEV
/etc/rc.d/boot.rootfsck: echo "fsck succeed, but reboot is required."
/etc/rc.d/boot.rootfsck: echo "fsck failed. Please repair manually and reboot. The root"
/etc/rc.d/boot.rootfsck: test $FSCK_RETURN -gt 0 && > /fsck_corrected_errors
/etc/rc.d/boot.rootfsck: echo '*** ERROR! Cannot fsck because root is not read-only!'
/etc/rc.d/boot.rootfsck: echo -n "Checking if boot.rootfsck has run"
/etc/rc.d/boot.swap:# Required-Start: boot.rootfsck $local_fs
/etc/rc.d/rc: if rc_cmdline forcefsck > /dev/null ; then
grep: /etc/rc.d/rc0.d: Is a directory
grep: /etc/rc.d/rc1.d: Is a directory
grep: /etc/rc.d/rc2.d: Is a directory
grep: /etc/rc.d/rc3.d: Is a directory
grep: /etc/rc.d/rc4.d: Is a directory
grep: /etc/rc.d/rc5.d: Is a directory
grep: /etc/rc.d/rc6.d: Is a directory
grep: /etc/rc.d/rcS.d: Is a directory
I don't use any dm or md RAID, LVM or other fancy stuff, so my most likely culprit would be /etc/rc.d/boot.rootfsck if I was having that problem. Your results may vary. The first step is finding the exact place where your boot scripts are stored.
going with daemon's post I looked at mv but wasnt sure would rsync be better than cp? or am i thinking to much?
I use rsync to update large nested directory structures with lots of files, like my `music' and `Pictures' directories; that's where it works best. You can use rsync to replace cp if you're copying over a network that's prone timeouts or anything that would interrupt a file copy, like over the Internet with cheap ADSL connections at both ends. The difference with rsync is that it checksums each file to ensure that it's the same at both ends. If you value your data, doing things like "copy then delete" and using rsync instead of cp can help protect your files so they don't get lost or corrupted when you move them.
i would rather hotspot hop until i find a free spot that is much faster (one is only four miles from me, another at Target, and so on) than waste eight hours and be unable to do anything else online during that time. and if the download fails....like it did on my Android phone (angry still as i was 98% done! now it just says 'download interrupted cannot resume')
That's a situation where rsync can be a big help. If you lose your connection at 98%, rsync will pick up right where it left off. You'd need shell access to the host that you're connecting to (or rsyncd running), which isn't always the case. However some public download servers offer the rsync protocol. Alternately you can use a FTP client with a "resume" mode.
Speed that download was via Android 2.3 using Dolphin Browser. it was not running from a *Nix box. for some reason if the phone phantom booted (twice it did that the first two attempts, the 3G keeps causing a random kernel panic, never happens in EDGE) or if the download was failed in any way (i live in a fringe 3G area, so it often drops out a lot) it would never resume. last i recalled, that is the whole point of a download manager--to resume failed or interrupted downloads. but i am not sure why it would not let me restart the failed download. it had the progress still stuck a 98% but Android would always say the same error--'Download interrupted. it cannot be resumed'
back to the fsck issue, when i looked in my /etc/fstab file, it did have a 'errors--remount ro' in the spot where it normally has the permissions (the 0 and 1) and since i was stuck in read-only i couldn't edit or save the file, just look at it. i couldn't even change the permissions manually. i had to dump into a recovery shell and do a fsck -C /dev/sda1 to scan and fix the errors before it would turn me loose and allow read/write access. the AUTO fsck that begins during system boot is bullocked. it is fine if the system is clean, but if it tries an auto-scan due to that 'dirty' bit it will freeze. i have to still hit C for cancel, and M for manual recovery to scan the system properly if the issue comes up again. if only there was some way to have it automatically dump me into a recovery shell when it is dirty again so it would not need lightning reflexes to catch the specific parts where i must hit C and M (it is a very fast boot). i mean Linux has for years, since my experience in BasicLinux, auto dumped into recovery shell if there was anything wrong and fsck saw a problem.
Last edited by nickdalzell; January 22nd, 2013 at 04:00 PM.
Speed that download was via Android 2.3 using Dolphin Browser. it was not running from a *Nix box.
Yes it was! Android runs on Linux, and there is an rsync client for Android. I have it on my phone and tablets, and use it to get media files from my Linux box.
Quote:
i live in a fringe 3G area, so it often drops out a lot
You might want to think about getting a cellular booster amplifier for your phone. Wilson Electronics makes a nice, inexpensive "cradle" model that works with any phone (doesn't require an antenna jack) and gives close to the maximum legal power to the antenna. The base model (there are 4G versions too) costs less than $100US, and if you spend a little extra for a good rooftop antenna and low loss RF cable to connect the two, you can see the fastest data rates possible. Although it's a mobile device, it can be used at home with a 12V DC power supply.
Quote:
back to the fsck issue, when i looked in my /etc/fstab file, it did have a 'errors--remount ro' in the spot where it normally has the permissions (the 0 and 1) and since i was stuck in read-only i couldn't edit or save the file, just look at it.
What you need is one of the several excellent rescue discs that can be used to boot the computer without using the Linux system installed on your HD. I prefer SystemRescueCD, but others like Parted Magic and even Knoppix (a full live distro) will work well. Some distros' install disc can also be booted into a recovery mode as well.
Just boot from the disc (or USB drive with the ISO image) and use the fsck on the boot disc to check, and hopefully repair the affected filesystem(s). Once you can run fsck without any errors, you can then mount the filesystem with /etc in it, and edit /etc/fstab (or copy fstab.bak to fstab if there is one) to restore normal operation.
If the filesystem has been damaged enough to corrupt crucial files, you may need to reinstall the OS. I've done this plenty of times (with my /home directory on a separate partition, of course) with no problems, but it can't hurt to make a backup. The rescue disk will allow you to copy your user directory to a spare disk or computer.
my last experience with KDE was back in Red Hat 7 on a bought disk set of the OS with every package possible at time of release (7 discs plus two for source code!) and it was downright primitive. other than the KDE theme and gear icons it looked indistinguishable from the GNOME it also came packaged with. it resembled the Default look of KDE 4. Windows 95-ish. the cool stuff is in the widgets and other myriad of desktop options. i installed Cairo Dock to replace the panel, use a few widgets, and went all out on the Compiz stuff. Being the Star Trek nut i themed Firefox in LCARS, installed an LCARS program to play with/use as screensaver, changed the default event sounds to Majel Barrett Roddenberry's computer voice samples, etc. there was also an LCARS-themed KDE load screen
Just installed fedora 18 atm i'm running kde. Havent ran that in for ever just thought I would check it out.
I hadn't run Fedora in a long time, but gave it a try in 2011 just for fun. I believe it was Fedora 16 (I have the disc around here somewhere, but I'm too lazy to look for it). I liked it, a lot, but ended up going back to my usual choice, Kubuntu, for that particular laptop.
The Following User Says Thank You to MoodyBlues For This Useful Post:
looks like i still got a bug with auto-fsck. even a proper shutdown and reboot will make it try and scan claiming it was umounted dirty--but it was a PROPER shutdown. there must be a bug in the kernel that keeps unmounting it wrong on shutdown. i just disabled the auto crap.
/dev/sda1 was uncleanly unmounted, check forced
*FREEZE* (just sits there indefinitely.)
i was able to edit the /etc/fstab file with a sudo gedit command. it would seem THIS distro still looks at that file to determine whether or not to scan.
Last edited by nickdalzell; January 23rd, 2013 at 08:53 AM.
looks like i still got a bug with auto-fsck. even a proper shutdown and reboot will make it try and scan claiming it was umounted dirty--but it was a PROPER shutdown. there must be a bug in the kernel that keeps unmounting it wrong on shutdown. i just disabled the auto crap.
/dev/sda1 was uncleanly unmounted, check forced
*FREEZE* (just sits there indefinitely.)
I've only skimmed most of the posts about this issue, but the conclusion I've come to is something I almost NEVER recommend. I believe Speed has already touched on it. I think doing a clean reinstall might be the solution here.
As I said, I almost never suggest this, as reinstalling the OS is just something that rarely needs to happen with *nix. But it seems that something(s) got so messed up somewhere along the way on your system that wiping it and starting fresh might be the best approach.
If you decide to do that, I'd recommend doing it with Kubuntu itself, rather than starting with Ubuntu and then adding KDE and Kubuntu components. Last time, you had already installed Ubuntu so it made sense to do it that way to try KDE, but now that you know you like it better, start fresh with Kubuntu from the get go.
isn't reinstalling something that happens to Microsoft Windows?
I will reiterate--it's been like this since day ONE. it is like my hard disk is not working correctly with it. i do know the hard disk is NOT bad. it is a brand new laptop. what i do know is going on, is when i shut the system down, it is NOT unmounting the file system, therefore it is no different than had i held down the power button and forced it off. so this causes the auto-fsck to think the system was unclean and force a disk check. so long as there are no errors, i can easily skip it and go about my day. if not, i can force a recovery shell, complete the scan myself and fix any file system errors. what i cannot do, however, is let it auto scan. if i let it auto-fsck it will freeze. every. time.
it took me hours to get what i got set up. it was a pain in the arse to install Star Trek Online. it was a 4.5 GB download. i am NOT about to mess that up by a reinstall.
isn't reinstalling something that happens to Microsoft Windows?
Yes. Which is why I said that reinstalling the OS is something that rarely needs to happen with *nix. But, sometimes, it may be called for. In my experience, that's most likely to occur right after it's been installed and, for whatever reason(s), something(s) got corrupted/screwed up. Instead of trying to find and fix the problems--which is certainly doable--it might just be easier to reinstall the OS.
Quote:
I will reiterate--it's been like this since day ONE. it is like my hard disk is not working correctly with it. i do know the hard disk is NOT bad. it is a brand new laptop. what i do know is going on, is when i shut the system down, it is NOT unmounting the file system, therefore it is no different than had i held down the power button and forced it off. so this causes the auto-fsck to think the system was unclean and force a disk check. so long as there are no errors, i can easily skip it and go about my day. if not, i can force a recovery shell, complete the scan myself and fix any file system errors. what i cannot do, however, is let it auto scan. if i let it auto-fsck it will freeze. every. time.
Which reinforces my feeling that a clean reinstall might be the remedy here.
Quote:
it took me hours to get what i got set up. it was a pain in the arse to install Star Trek Online. it was a 4.5 GB download. i am NOT about to mess that up by a reinstall.
How did you partition the disk? If you did it in the 'best practices' sort of way, you should have the OS installed on its own, separate / (root) partition, a separate partition for /home, some swap space, and perhaps some other partition(s) for data, downloads, whatever. If you did it this way, and assuming that you have not placed downloaded files on your root partition (and, therefore, won't lose those when you format /), reinstalling the OS will not affect anything in your /home partition, including STO (which I believe you're running via wine, correct? so its files are in your ~/.wine directory). If you decide to reinstall, choose to format / and NOT format /home. That way you won't lose anything except system files and software you've installed, for example, via Synaptic, that placed the files in /usr/games, /usr/local, etc. But you can reinstall those later. Any data should be on your /home partition and will be safe.
(Of course I always recommend doing a backup first. Personally, I have never lost any data after decades of doing clean installs (which I prefer when upgrading), but no one can guarantee anything. Better safe than sorry.)
The Following 2 Users Say Thank You to MoodyBlues For This Useful Post:
Device(s): HTC EVO 4g(Retired), HTC EVO LTE(Retired), LG Lucid(Returned), Motorola Droid Razr M, Nexus 7
Carrier: GOD
Thanks: 3,752
Thanked 3,962 Times in 1,980 Posts
When ever i install a system i never format /home however i do have a current backup of it. The only time I lost something it was my own fault. Or it was a Windows thing and we wont go there. However unlike windows reinstalling a Linux doesn't have to be a headache it can be done fairly simply and guess what not formatting home will keep all your settings
i got the entire disk devoted to root. or /. there is no swap space as i got 4GB of RAM. i never run out of RAM and while it is running it works well. great, in fact. it is specifically how it tries unmounting the system when shut down. or some feature say maybe APM is preventing auto-fsck to work correctly. i found workarounds but no real fixes. nothing else is wrong with it.
could having the file system set to Ext2 vs Ext4 be a problem? honestly i have never noticed a single difference one way or the other in terms of stability or performance
for the record, this bug happens on both laptops. one running Kubuntu, (just got done installing it at home), and my STO computer running Ubuntu w/ Kubutu-desktop. both have auto-fsck failures. both won't unmount the root file system properly on shutdown. so the issue is not with one computer. i installed via two different disks too. when i view the kernel output on shutdown (i prefer verbose, it's fun watching it scroll on the screen) it never even mentions unmounting filesystems. not once. it does have a 'fail' flag on some kernel msg daemon but that is it
both have SATA hard drives, perhaps i got them configured for IDE or something?
Last edited by nickdalzell; January 23rd, 2013 at 09:17 PM.
282Likes
Back in the DOS 8.3 (file names limited to 8 characters, plus a MANDATORY dot-plus-3-character extension), in UNIX we had no such ridiculous limitations.
With their non-standards-compliant, so-called newsgroup clients--that did things like put background images in their posts...or attempted to, anyway...those of us using real newsgroup clients just saw a bunch of m$ garbage.
is to keep the source file(s) until you're sure that they all made it to their destination intact. So instead of mv, you might do a cp, check to see that the files are there, and then rm the source files. This takes longer and requires more disk space (which may be a problem if your reason to move the files is a full disk). But it's the safest way to do it.
How 'bout we mutually agree to drop this, and get back to discussing Linux topics in this thread? 
