|
Suspicious Dish posted:Want me to make a new thread then? "The Linux Questions Thread: Red Hat is mostly two words"
|
#
?
Feb 5, 2014 23:30
|
|
|
|
| # ? May 30, 2024 15:52 |
|
|
Bhodi posted:If you can be neutral! I definitely agree, it needs an update. And, since you're offering... The Linux questions Thread: Want Linux? Use Fedora. The Linux Questions Thread: The GUI is for a pretty terminal.
|
|
#
?
Feb 5, 2014 23:34
|
|
|
When kickstarting, there's an option (sshpw setting) that lets users ssh in to the currently installing host. However, the actually install process can only be seen by VNCing in because the install runs in GUI mode. If I disable GUI mode install (text setting), can I watch the install when I ssh in? Or will I be dropped to a prompt as usual? I'd like to be able to ssh in and debug an install without having to connect via VNC.
|
|
#
?
Feb 6, 2014 02:33
|
|
|
Ninja Rope posted:When kickstarting, there's an option (sshpw setting) that lets users ssh in to the currently installing host. However, the actually install process can only be seen by VNCing in because the install runs in GUI mode. If I disable GUI mode install (text setting), can I watch the install when I ssh in? Or will I be dropped to a prompt as usual? Pipe all the textmode output to a log file, tail -f the log file.
|
|
#
?
Feb 6, 2014 03:50
|
|
|
RFC2324 posted:Pipe all the textmode output to a log file, tail -f the log file. What? Beyond anaconda being curses and generally not piping well, and kickstarts being impossible to pipe without writing another kickstart to build another CD where anaconda has redirection (anaconda itself has a lot going on in VTs, but mostly it passes args from cmdline), it already logs in /tmp. The point of sshpw is that you can easily log in and grab logs in case something goes wrong, poke around the system, and generally troubleshoot. It's not particularly useful otherwise.
|
|
#
?
Feb 6, 2014 04:37
|
|
|
RFC2324 posted:Pipe all the textmode output to a log file, tail -f the log file. I don't know how to do that, I think kickstart starts anaconda(?) without me being able to modify how it's run. Also, I assume anaconda might prompt for questions and I wouldn't be able to answer them if I'm just dumping to a file.
|
|
#
?
Feb 6, 2014 05:30
|
|
|
Ninja Rope posted:I don't know how to do that, I think kickstart starts anaconda(?) without me being able to modify how it's run. Also, I assume anaconda might prompt for questions and I wouldn't be able to answer them if I'm just dumping to a file. Kickstarts are all or nothing and won't prompt. You can't easily modify how anaconda runs without creating a custom image. Anaconda isn't text output and can't be legibly pipes to a file any more than mc can.
|
|
#
?
Feb 6, 2014 05:39
|
|
|
evol262 posted:Kickstarts are all or nothing and won't prompt. It definitely does prompt, at least to retry if it fails to fetch some packages (and the documentation states it will prompt for more than that if necessary). The kickstart docs say that the "text" option runs anaconda in text mode, but I'm not sure how that helps if I can't view what the installer is doing. Maybe this option isn't used very often?
|
|
#
?
Feb 6, 2014 06:27
|
|
|
apt-get dist-upgrade is still a marvel to me. Just finished upgrading from debian squeeze to wheezy and despite the face my system was in a disaray of contradictory repositories, pinned and generally messed up packages, it upgraded perfectly , except for a slight hickup with java (gently caress you oracle!). I shouldn't be surprised though, many years ago on a bet I apt-get dist-updated from debian to ubuntu, and it worked flawlessly. Then a month later I dist-upgraded back. The magics still there. 
|
|
#
?
Feb 6, 2014 10:58
|
|
|
So I have a Lenovo laptop that I have Ubuntu server 11.10, running a clonezilla server. I can plug in thumb drives all day long and they are recognized and I can mount them. If I plug in my WD usb hard drive I get nothing, the system has no idea anything has been plugged in. I was thinking it might have been a power issue but if I boot to say Parted Magic and plug it in its recognized and I can mount it. Any thoughts on what component I am missing which is keeping me from using it under 11.10?
|
|
#
?
Feb 6, 2014 13:52
|
|
|
niss posted:So I have a Lenovo laptop that I have Ubuntu server 11.10, running a clonezilla server. I can plug in thumb drives all day long and they are recognized and I can mount them. If I plug in my WD usb hard drive I get nothing, the system has no idea anything has been plugged in. I was thinking it might have been a power issue but if I boot to say Parted Magic and plug it in its recognized and I can mount it. Any thoughts on what component I am missing which is keeping me from using it under 11.10? What shows up in /var/syslog when you plug the drive in?
|
|
#
?
Feb 6, 2014 14:38
|
|
|
Ninja Rope posted:It definitely does prompt, at least to retry if it fails to fetch some packages (and the documentation states it will prompt for more than that if necessary). The kickstart docs say that the "text" option runs anaconda in text mode, but I'm not sure how that helps if I can't view what the installer is doing. Maybe this option isn't used very often? "text" means "not GUI" in this case, but it's curses. It's been sort of deprecated since rhel5, in the sense that new anaconda features don't get added to the text ui, but it works fine for kickstarts. And, yes, kickstarted installs can prompt you to retry/cancel when it hits an error, but this doesn't mean you should plan on that. Host a local yum repo. The whole point of kickstarting is that it's fire-and-forget. If you're waiting to hit keys, there's no point to it. You should be able to pxe boot an image or boot a CD, walk away, and come back in 5 minutes to have a system ready to log into.
|
|
#
?
Feb 6, 2014 15:15
|
|
|
Bob Morales posted:What shows up in /var/syslog when you plug the drive in? Nothing shows up in /var/log/syslog However if I plug the thumb drive in I get data related to the OS seeing it. The really strange thing is this server is a clone from a desktop system which I can insert the USB hard drive with zero issues.
|
|
#
?
Feb 6, 2014 15:56
|
|
|
niss posted:Nothing shows up in /var/log/syslog Check dmesg, not /var/log/syslog
|
|
#
?
Feb 6, 2014 16:07
|
|
|
Speaking of Lenovo, a word of warning. Installing Linux on T540p/L540/W540 Thinkpads bricks the mobo. All hail lovely UEFI implementations!
|
|
#
?
Feb 6, 2014 16:17
|
|
|
evol262 posted:Check dmesg, not /var/log/syslog I got nothing related to that hard drive 
|
|
#
?
Feb 6, 2014 16:21
|
|
|
niss posted:I got nothing related to that hard drive If nothing's showing up in dmesg when you plug it in, it's not even appearing on the bus. Give more power to the drive.
|
|
#
?
Feb 6, 2014 16:44
|
|
|
evol262 posted:If nothing's showing up in dmesg when you plug it in, it's not even appearing on the bus. Give more power to the drive. That was my initial thought as well, only problem is it is USB powered, no external power plug. That still doesn't' seem to explain why it would work under a bootable live linux environment.
|
|
#
?
Feb 6, 2014 16:55
|
|
|
niss posted:That was my initial thought as well, only problem is it is USB powered, no external power plug. That still doesn't' seem to explain why it would work under a bootable live linux environment. Same USB ports? Those 2.5" USB caddies often draw power from 2 USB ports because .5A isn't enough (if it's USB2). Is it a bootable environment on the same hardware? Do you see a power draw from the port with lsusb?
|
|
#
?
Feb 6, 2014 17:03
|
|
|
evol262 posted:Same USB ports? exact same everything. Just tested lsusb and nothing. I know the drive is working cause I just booted the system to parted magic to copy an image file from the system to the drive. Its really weird.
|
|
#
?
Feb 6, 2014 17:10
|
|
|
evol262 posted:Those 2.5" USB caddies often draw power from 2 USB ports because .5A isn't enough (if it's USB2). That's very rare. I've only seen USB optical drives that want 2 USB cables for power.
|
|
#
?
Feb 6, 2014 17:18
|
|
|
niss posted:exact same everything. Just tested lsusb and nothing. I know the drive is working cause I just booted the system to parted magic to copy an image file from the system to the drive. Its really weird. Even if the device isn't there, lsusb -v should show amperage used by port. Is there any? Bob Morales posted:That's very rare. I've only seen USB optical drives that want 2 USB cables for power. It's so common that 75% of the enclosures on first page of Amazon's storefront have them. The "one mini-USB, two USB A" cable is designed so you can plug in both USB A ends to different ports and actually draw 1A. There's only one USB port on the enclosure, but it draws from two ports.
|
|
#
?
Feb 6, 2014 18:49
|
|
|
evol262 posted:Even if the device isn't there, lsusb -v should show amperage used by port. Is there any? So I issued the lsusb -v command, it scrolled a ton of data across the screen, then did a fdisk -l and lo and behold it now shows up. Rebooted the system plugged it in and once again nothing. Issued lsusb -v and it once again shows up. Strange, but atleast I know what I need to do on the system to access the drive. Thanks
|
|
#
?
Feb 6, 2014 18:59
|
|
|
evol262 posted:"text" means "not GUI" in this case, but it's curses. It's been sort of deprecated since rhel5, in the sense that new anaconda features don't get added to the text ui, but it works fine for kickstarts. Yes, I understand that, but thank you. The issue I've been having is when kickstart hangs for some reason (for example, when the repo is inaccessible because the network is misconfigured). In that case anaconda pauses and pops up a "retry?" box you have to click on. If the install is run headless and I don't watch VNC, the kickstart appears to hang forever and I don't know why. I'd like a way to debug a running kickstart and see what, if anything, is going wrong. It seems like the only way to do that now is VNC, and I was hoping there was a non-GUI way of doing that. I'll just live with VNC for now.
|
|
#
?
Feb 6, 2014 21:51
|
|
|
Ninja Rope posted:Yes, I understand that, but thank you. The issue I've been having is when kickstart hangs for some reason (for example, when the repo is inaccessible because the network is misconfigured). In that case anaconda pauses and pops up a "retry?" box you have to click on. If the install is run headless and I don't watch VNC, the kickstart appears to hang forever and I don't know why. I'd like a way to debug a running kickstart and see what, if anything, is going wrong. It seems like the only way to do that now is VNC, and I was hoping there was a non-GUI way of doing that. This is the exact use of sshpw. Just that Anaconda logs in /tmp, and you should watch there (anaconda.log, storage.log, etc..)… grab the error, fix the kickstart, and bounce the server until it installs. Why click "retry" when your unattended install needs to be attended? Grab the error out of /tmp, fix, and reboot. That's why sshpw exists. Also, anaconda --ksfile=... works if you ssh in and wanna debug without rebooting.
|
|
#
?
Feb 7, 2014 02:38
|
|
|
evol262 posted:This is the exact use of sshpw. Just that Anaconda logs in /tmp, and you should watch there (anaconda.log, storage.log, etc..)… grab the error, fix the kickstart, and bounce the server until it installs. Why click "retry" when your unattended install needs to be attended? Grab the error out of /tmp, fix, and reboot. That's why sshpw exists. Do I need to set "text" to get the logs in /tmp? I looked there before and I didn't see anything, but maybe I'm blind...
|
|
#
?
Feb 7, 2014 02:47
|
|
|
Ninja Rope posted:Do I need to set "text" to get the logs in /tmp? I looked there before and I didn't see anything, but maybe I'm blind... Nope. GUI (regular) anaconda should spin up multiple VTs with various logs and log to /tmp as normal.
|
|
#
?
Feb 7, 2014 03:38
|
|
|
http://aceattorney.sparklin.org/jeu.php?id_proces=57684 linux
|
|
#
?
Feb 7, 2014 18:07
|
|
|
Prepare for what is likely an incredibly stupid question. My laptop has both integrated graphics on it as well as a better but more resource-intensive Nvidia card. Except right now, I'm not sure I'm using either. I'm currently running Linux Mint with MATE. There were some default nvidia drivers that showed up under the Driver Manager, but I installed nvidia-current and they don't show up anymore, and Nvidia X Server Settings always just tells me "You do not appear to be using the NVIDIA X driver. Please edit your X configuration file (just run `nvidia-xconfig` as root), and restart the X server.". (I have done that several times). My main suspicion that something is messed up is that if I try to play any videos or games with hardware acceleration, my computer immediately crashes. I feel like I picked the worst possible laptop to try and use linux on.
|
|
#
?
Feb 10, 2014 21:55
|
|
|
bomblol posted:Prepare for what is likely an incredibly stupid question. My laptop has both integrated graphics on it as well as a better but more resource-intensive Nvidia card. Except right now, I'm not sure I'm using either. I'm currently running Linux Mint with MATE. There were some default nvidia drivers that showed up under the Driver Manager, but I installed nvidia-current and they don't show up anymore, and Nvidia X Server Settings always just tells me "You do not appear to be using the NVIDIA X driver. Please edit your X configuration file (just run `nvidia-xconfig` as root), and restart the X server.". (I have done that several times). My main suspicion that something is messed up is that if I try to play any videos or games with hardware acceleration, my computer immediately crashes. You're almost certainly using the intel drivers, but they're dying on nVidia's opengl stack because you installed the binary drivers. Output of 'lsmod' and 'glxinfo |grep vendor', please. If this is Optimus (I'm guessing it is), nVidia sucks at this, and they haven't bothered with Linux drivers, so the community has sort-of reversed them. You want Bumblebee. I don't use Mint or Optimus, but this looks reasonably good for intstall instructions.
|
|
#
?
Feb 10, 2014 22:11
|
|
|
evol262 posted:You're almost certainly using the intel drivers, but they're dying on nVidia's opengl stack because you installed the binary drivers. http://pastebin.com/nQWYP0i0 is my lsmod, and code:
|
|
#
?
Feb 11, 2014 00:41
|
|
|
I have a machine that was being backed up using rdiff-backup and it suddenly stopped working and it appears that the issue is that the entire file system was being backed up and the contents of /run/udev/links/ are understandably not able to be processed properly by rdiff-backup. That directory doesn't actually need to be backed up so I tried omitting it from the backup but rdiff-backup still attempts to read those increment files from the remote share. When I try to remove those files directly from the remote share, I'm unable to do so. I've tried escaping the backslashes in the directory names as well as putting the directories in quotes with no success. I also cannot delete them from the remote server. Any ideas on how to delete these?pre:$ pwd /mnt/backup2/rdiff-backup-data/increments/run/udev/links $ ls disk\x2fby-id\x2fata-HL-DT-ST_DVD+_-RW_GHA2N_K4ID48D3549 disk\x2fby-id\x2fata-ST2000DM001-1CH164_S1E20E67 disk\x2fby-id\x2fata-ST2000DM001-1CH164_S1E20E67-part1 disk\x2fby-id\x2fata-ST2000DM001-1CH164_S1E20E67-part2 ...
|
|
#
?
Feb 11, 2014 15:33
|
|
|
Both drivers are in kernel. lspci ? Is this Optimus? Modern Pragmatist posted:I have a machine that was being backed up using rdiff-backup and it suddenly stopped working and it appears that the issue is that the entire file system was being backed up and the contents of /run/udev/links/ are understandably not able to be processed properly by rdiff-backup. That directory doesn't actually need to be backed up so I tried omitting it from the backup but rdiff-backup still attempts to read those increment files from the remote share. When I try to remove those files directly from the remote share, I'm unable to do so. I've tried escaping the backslashes in the directory names as well as putting the directories in quotes with no success. I also cannot delete them from the remote server. Any ideas on how to delete these? What's the message when you try to delete them?
|
|
#
?
Feb 11, 2014 16:18
|
|
|
evol262 posted:What's the message when you try to delete them? pre:$ rm -r "disk\x2fby-id\x2fata-HL-DT-ST_DVD+_-RW_GHA2N_K4ID48D3549" rm: cannot remove `disk\\x2fby-id\\x2fata-HL-DT-ST_DVD+_-RW_GHA2N_K4ID48D3549': Invalid argument
|
|
#
?
Feb 11, 2014 19:37
|
|
|
lsof show anything? Try rm -i * and choose your file.
|
|
#
?
Feb 11, 2014 19:43
|
|
|
hackedaccount posted:lsof show anything? pre:# lsof | grep /mnt/backup2 bash 11283 user cwd DIR 0,25 0 9895851196697 /mnt/backup2/rdiff-backup-data/increments/run/udev/links lsof 27800 user cwd DIR 0,25 0 9895851196697 /mnt/backup2/rdiff-backup-data/increments/run/udev/links grep 27801 user cwd DIR 0,25 0 9895851196697 /mnt/backup2/rdiff-backup-data/increments/run/udev/links lsof 27802 user cwd DIR 0,25 0 9895851196697 /mnt/backup2/rdiff-backup-data/increments/run/udev/links pre:# rm -ri disk* rm: descend into write-protected directory `disk\\x2fby-id\\x2fata-HL-DT-ST_DVD+_-RW_GHA2N_K4ID48D3549'? y rm: cannot remove `disk\\x2fby-id\\x2fata-HL-DT-ST_DVD+_-RW_GHA2N_K4ID48D3549': Invalid argument rm: cannot remove `disk\\x2fby-id\\x2fata-ST2000DM001-1CH164_S1E20E67': Invalid argument rm: cannot remove `disk\\x2fby-id\\x2fata-ST2000DM001-1CH164_S1E20E67-part1': Invalid argument rm: cannot remove `disk\\x2fby-id\\x2fata-ST2000DM001-1CH164_S1E20E67-part2': Invalid argument ...
|
|
#
?
Feb 11, 2014 20:07
|
|
|
Modern Pragmatist posted:
This is probably bash escapes being dumb. python -c "import os; import glob; [os.unlink(i) for i in glob.glob(u'disk*')]"
|
|
#
?
Feb 11, 2014 20:20
|
|
|
Why are you trying to delete stuff in /dev? All that stuff is managed by udev and accompanying udev rule files.
|
|
#
?
Feb 11, 2014 21:00
|
|
|
waffle iron posted:Why are you trying to delete stuff in /dev? All that stuff is managed by udev and accompanying udev rule files. It's not on the system. It's rsnapshotted somewhere else.
|
|
#
?
Feb 11, 2014 21:41
|
|
|
|
| # ? May 30, 2024 15:52 |
|
|
waffle iron posted:Why are you trying to delete stuff in /dev? All that stuff is managed by udev and accompanying udev rule files. Not in /dev, but in /mnt/backup2/..., which is an accidentally-made backup of the real /dev as it existed in one point of time, on a CIFS filesystem. Modern Pragmatist just wants to get rid of the useless backup, and it is turning out to be difficult. I guess the problem might be the \x2f, which seems to be a hexadecimal escape of... the forward slash, or "/". The thing that separates Unix-style pathname elements from each other, and thus cannot really be part of any Unix-style pathname element. But on a CIFS filesystem, the separator is the backslash "\", so a forward slash is just another character. Either the CIFS driver or the VFS layer translates the embedded forward slash to \x2f but does not translate it back at the right spot, or rdiff-backup did the translation when creating the directories and something in CIFS/VFS is too eager in applying the reverse translation. My first suggestion would be to bypass the entire POSIX/VFS hierarchy and try using smbclient to delete the offending files/directories. Smbclient won't need to go through the VFS translation layers, and will construct CIFS network protocol commands directly, so it should be able to handle a forward slash in a remote file/dir name just fine. If the filenames (as seen by the smbclient) contain actual forward slashes, think about sending a bug report to the CIFS developers. (Of course, if your kernel/CIFS software versions are far from the latest ones, it might already be fixed.)
|
|
#
?
Feb 11, 2014 21:47
|
|