|
MargotK posted:Again, thanks a lot!
|
# ? Mar 9, 2009 02:34 |
|
|
# ? May 10, 2024 03:21 |
|
I have a raid 5 using mdadm in my home storage computer. Right now it's set up with 3x 500 gig drives as raid devices and a spare 500 is a hotswap. I've nearly filled the storage pool and would like to add in the hotwap drive to the storage pool. I read the man page but am unsure of what I do exactly. I've figured that I'd have to use mdadm with the grow command but am unsure of which drives to point to how in the proper syntax. Any help would be appreciated as I feel silly for not being able to figure it out. And to avoid any argument I have my data backed up elsewhere and I know raid isn't a form of backup etc. Here's the printout of: sudo mdadm --detail /dev/md0 code:
|
# ? Mar 9, 2009 19:48 |
|
Slow is Fast posted:I've figured that I'd have to use mdadm with the grow command but am unsure of which drives to point to how in the proper syntax. mdadm --grow --raid-devices 4 /dev/md0
|
# ? Mar 9, 2009 20:48 |
|
ShoulderDaemon posted:mdadm --grow --raid-devices 4 /dev/md0 That's it! I suck at understanding man pages. I'll give it a go and report back. Edit: raid looks good, it's flagged as "clean, rebuilding" in webmin and I hear the drives doing their magic. Now all I have to do is poke around and figure out how to get email notifications sent to me if bad things happen. Slow is Fast fucked around with this message at 21:24 on Mar 9, 2009 |
# ? Mar 9, 2009 21:00 |
|
Here's something that's bugging me. I'm trying to raise the limit of open files on a Linux box (I'm running a distributed computing network on it, 1024 open files really isn't enough.) Since I'm going to have programs SSH'ing in and running other programs, it rather needs to be automatic. "ulimit -n" reports 1024, and that isn't enough, so I went about trying to change it. Added to /etc/security/limits.conf: code:
code:
I ran into something online which mentioned that there was an SSH bug such that "UsePrivilegeSeparation yes" in the file made it ignore the limits.conf file. I changed it to "no" and it's still 1024. There have been multiple reboots through all of this, so I don't have just stale servers running or anything All of this is being done via SSH, since I can't really log in directly. Annoyingly, I have a VM with what I thought was the same setup, and it works just fine ("ulimit -n" gives 32768) - the only known difference is that it's Ubuntu 8.10 Desktop, and the new one is Ubuntu 8.10 Server. I'm out of ideas on how to even debug this one, I can't find any info on how to track where the ulimit is coming from. Suggestions? ZorbaTHut fucked around with this message at 07:16 on Mar 10, 2009 |
# ? Mar 10, 2009 07:05 |
|
Slow is Fast posted:That's it! I suck at understanding man pages. I'll give it a go and report back. Heres a handy tip to force it to speed up rebuilding if you are impatient: http://www.ducea.com/2006/06/25/increase-the-speed-of-linux-software-raid-reconstruction/
|
# ? Mar 10, 2009 08:04 |
|
ZorbaTHut posted:Here's something that's bugging me. * soft nofile 32768 * hard nofile 32768 Make sure "UsePAM" is set to yes in /etc/ssh/sshd.conf. (limits.conf is invoked by pam_limits.so) If nothing works, try putting "ulimit -n 32768" into the .bashrc for your users.
|
# ? Mar 10, 2009 13:24 |
|
My raid rebuilt itself and its showing the other drive as being added. The only problem I'm having is that my other computers aren't able to use the new space. Windows mapping the network drive still reports it as only being 1tb and not 1.5 like it should be with the new drive. I deleted the mount point and tried redoing it to see if that was the problem. I'm not sure where to go from here.
|
# ? Mar 10, 2009 17:25 |
|
Slow is Fast posted:My raid rebuilt itself and its showing the other drive as being added. The only problem I'm having is that my other computers aren't able to use the new space. Windows mapping the network drive still reports it as only being 1tb and not 1.5 like it should be with the new drive. I deleted the mount point and tried redoing it to see if that was the problem. I'm not sure where to go from here. You've expanded the size of the underlying device, but not the filesystem on top of it. What filesystem is it?
|
# ? Mar 10, 2009 20:40 |
|
chryst posted:First try this in limits.conf, just in case the "-" isn't working for some reason. Tried these without luck, I finally managed to track it down - for some bizarre reason the * wasn't working. I had to specify the user name in limits.conf. zorba - nofile 32768 No, I can't explain this either. Thanks for the try, though
|
# ? Mar 10, 2009 21:03 |
|
I'm using Ubuntu 8.10 and have an annoying mimetype issue. How can I get jpg and JPG to show up as the same? The system does not treat jpg and JPG's as the same file type. It has created an entry for JPG, but I was hoping to just add it as one of the extensions. The single entry does not have all the nautilus support etc that jpg does. My camera uses all upper case when saving pictures, and I haven't found a way to change that.
|
# ? Mar 10, 2009 21:11 |
|
ShoulderDaemon posted:You've expanded the size of the underlying device, but not the filesystem on top of it. What filesystem is it? Oh poo poo you're right. I'm using xfs. The man page for xfs. http://linux.die.net/man/8/xfs_info Looks like it's just: xfs_growfs -d /dev/md0 correct? Should I unmount it from /mnt/storage as well? Thanks! It worked. vvvv Slow is Fast fucked around with this message at 00:10 on Mar 11, 2009 |
# ? Mar 10, 2009 21:55 |
|
Slow is Fast posted:Oh poo poo you're right. I'm using xfs. XFS needs to be mounted in order to be grown. If it's mounted at /mnt/storage, then xfs_growfs /mnt/storage should be all you need.
|
# ? Mar 10, 2009 23:24 |
|
ZorbaTHut posted:Tried these without luck, I finally managed to track it down - for some bizarre reason the * wasn't working. I had to specify the user name in limits.conf.
|
# ? Mar 11, 2009 18:17 |
|
Does anybody know of a good live CD that supports NTFS read/write? I'm trying to save some files of a dieing NTFS drive by offloading them to a working NTFS drive. For dded fun, all I have to boot off of currently is a Ubuntu 7.10 Live CD with no hard drive (though with a 4GB USB stick and I can install a second DVD burner, but if I can get away with only one DVD drive that'd save me some work.)
|
# ? Mar 12, 2009 07:48 |
|
Magicmat posted:Does anybody know of a good live CD that supports NTFS read/write? I'm trying to save some files of a dieing NTFS drive by offloading them to a working NTFS drive. For dded fun, all I have to boot off of currently is a Ubuntu 7.10 Live CD with no hard drive (though with a 4GB USB stick and I can install a second DVD burner, but if I can get away with only one DVD drive that'd save me some work.) RIP has ntfs3g and some cool tools: http://rip.7bf.de/current/ Theres also NTFS-3g on the partedmagic boot cd I think.
|
# ? Mar 12, 2009 11:42 |
|
Ubuntu 8.10 Live CDs use NTFS 3g. I'm not 100% positive, but I know I was doing read/write to my NTFS drive off one last week, and I'm fairly sure that's a tell-tale sign.
|
# ? Mar 12, 2009 15:10 |
|
LiquidRain posted:Ubuntu 8.10 Live CDs use NTFS 3g. I'm not 100% positive, but I know I was doing read/write to my NTFS drive off one last week, and I'm fairly sure that's a tell-tale sign.
|
# ? Mar 12, 2009 15:30 |
|
jegHegy posted:I think I recall the live CD automounting my USB HDD formatted to NTFS since at least 7.10, maybe 7.04. Anyway, I think it should be in both fedora and ubuntus lates livecds. I know it's in the installed version of fedora.
|
# ? Mar 12, 2009 15:38 |
|
Even if ntfs 3g isn't on an Ubuntu livecd by default, you can install it via the package manager.
|
# ? Mar 12, 2009 18:31 |
|
Awesome, Ubuntu 8.10 worked out-of-the-box perfectly! Score another one for Ubuntu, I guess. Now for an even better question: I need to run the Samsung hard drive diagnostic utility, esTool. Problem is, it only runs off a DOS boot disk. So how do I create a DOS boot disk from within Linux? I'd prefer to get it running off my USB stick, but I'll settle for burning an entire CD for a 2 MB utility, I guess. Samsung, as some sort of sick joke, provides two versions of esTool on their page, one for a floppy and one for a CD-ROM. The floppy version is just a zip containing estool.exe; the CD version, however, is zip file containing an ISO. Now, if you were expecting an ISO -- a file format meant to contain an exact replica of a disk -- to be able to be burned to a disk and you're good to go, oh ho ho, jokes on you: the ISO just contains estool.exe and a batch file called run.bat containing only the line "estool.exe"
|
# ? Mar 12, 2009 23:08 |
|
This page explains where to get and how to add stuff to DOS boot disks in Linux
|
# ? Mar 13, 2009 03:23 |
|
on an ubuntu system, say i have a broken installation of something (samba in this case, i deleted /etc/samba), how can i get apt-get install samba to run again, and re-write it's configs and such?
|
# ? Mar 13, 2009 07:00 |
|
Something Awesome posted:on an ubuntu system, say i have a broken installation of something (samba in this case, i deleted /etc/samba), how can i get apt-get install samba to run again, and re-write it's configs and such? code:
|
# ? Mar 13, 2009 08:03 |
|
I want to update Pylint to a newer version on my laptop running ubuntu hardy. But after adding the repo to synaptic, it doesn't pick up the more recent version. Does that mean hardy can't use anything past .13 or am I doing something wrong? I am using pylint with pydev and it does not handle with statements correctly, meaning modules using with lose all static checking and warnings.
|
# ? Mar 13, 2009 21:36 |
|
hlfrk414 posted:I want to update Pylint to a newer version on my laptop running ubuntu hardy. But after adding the repo to synaptic, it doesn't pick up the more recent version. Does that mean hardy can't use anything past .13 or am I doing something wrong? I am using pylint with pydev and it does not handle with statements correctly, meaning modules using with lose all static checking and warnings. Just making sure you did the basics, did you do a sudo apt-get update after you added the url?
|
# ? Mar 13, 2009 22:53 |
|
musclecoder posted:Just making sure you did the basics, did you do a sudo apt-get update after you added the url? Yup. Normally I can figure stuff like this out, but I don't know what to look for. I've added the repo along with the signed key, I've done similar before and when I've done it wrong it complains very loudly. Nothing in ubuntu should be relying on pylint, so no dependencies should prevent me from using the newest version. And pylint 1.4 and above support python 2.5, so that shouldn't be an issue. I can't see what I'm missing.
|
# ? Mar 13, 2009 23:20 |
|
hlfrk414 posted:I want to update Pylint to a newer version on my laptop running ubuntu hardy. But after adding the repo to synaptic, it doesn't pick up the more recent version. Does that mean hardy can't use anything past .13 or am I doing something wrong? I am using pylint with pydev and it does not handle with statements correctly, meaning modules using with lose all static checking and warnings. That repository does not appear to include pylint. What version of pylint do you need? It looks like it shouldn't be too hard for me to throw together a backport for you against hardy. Edit: If you have the hardy-updates apt source configured (which you should), then you should be able to install this .deb from jaunty which will bring you to version 0.15.2. If you need better than that, Debian has 0.16.0 which will probably install directly, but if not can be rather trivially repackaged for your version. ShoulderDaemon fucked around with this message at 23:55 on Mar 13, 2009 |
# ? Mar 13, 2009 23:49 |
|
Thanks, that's good enough for me. I totally forgot that I could get a deb, but I kinda question why that repository doesn't contain pylint. Ah well.
|
# ? Mar 14, 2009 00:03 |
|
hlfrk414 posted:Thanks, that's good enough for me. I totally forgot that I could get a deb, but I kinda question why that repository doesn't contain pylint. Ah well. Silly question, why are you wanting to install it via Apt anyways? I've found the best/easiest/least painful way to manage Python modules is via easy_install or pip, so that only my core Python itself (and `python-setuptools` for easy_install) are in Apt and the rest I install via easy_install/pip. In terms of being able to upgrade, I think those tools are capable of upgrading; in terms of removal, they can't, but use of virtualenv makes it (and management of Python libs overall) a whole lot easier. (Note that all this doesn't imply that doing it via Apt is bad per se, but I've been a Debian sysadmin and Python coder for years, and wrestled with the problem of Python libs in/out of Apt for a long time. Switching to the above setup has been amazing in terms of keeping things straight for me.)
|
# ? Mar 14, 2009 00:37 |
|
This has been strangely difficult to google, I have a remote ssh connection into my box that's been there since december 2008, I want to kill it. How? (preferably not killing the other connections coming in)
|
# ? Mar 15, 2009 20:57 |
|
NZAmoeba posted:This has been strangely difficult to google, I have a remote ssh connection into my box that's been there since december 2008, I want to kill it. How? (preferably not killing the other connections coming in) "ps aux | grep ssh" should give you a list of processes. I forget if sshd sets the username of connecting user. Edit: Maybe the output of "who" might list it. waffle iron fucked around with this message at 21:49 on Mar 15, 2009 |
# ? Mar 15, 2009 21:46 |
How do I update my time zone data in zoneinfo? I'm running Fedora Core 7. I tried yum update tzdata (which found an update and installed it, still in MST instead of MDT though).
|
|
# ? Mar 16, 2009 03:56 |
|
fletcher posted:How do I update my time zone data in zoneinfo? I'm running Fedora Core 7. I tried yum update tzdata (which found an update and installed it, still in MST instead of MDT though). http://rpm.pbone.net/index.php3/stat/4/idpl/11518506/com/tzdata-2009a-1.fc10.noarch.rpm.html waffle iron fucked around with this message at 04:57 on Mar 16, 2009 |
# ? Mar 16, 2009 04:42 |
waffle iron posted:Your best bet is to download a tzdata rpm from the oldest supported Fedora updates and install that. Fedora 9 should do the trick. tzdata is noarch anyway. Ok, got that installed. What do I do to update the time after that? code:
|
|
# ? Mar 16, 2009 04:55 |
|
fletcher posted:Ok, got that installed. What do I do to update the time after that?
|
# ? Mar 16, 2009 04:58 |
waffle iron posted:Either copy or symlink the right tzdata file to /etc/localtime to make sure it's the right thing. Yay, thanks! For future reference for the other idiots like me, this command: code:
code:
fletcher fucked around with this message at 05:19 on Mar 16, 2009 |
|
# ? Mar 16, 2009 05:16 |
|
NZAmoeba posted:This has been strangely difficult to google, I have a remote ssh connection into my box that's been there since december 2008, I want to kill it. How? (preferably not killing the other connections coming in) Running "lsof -i :22" would show you exactly which PIDs are active (on port 22), just kill the appropriate one.
|
# ? Mar 16, 2009 06:52 |
|
Another question! (Thanks for the prior help, issue solved) I'm trying to install rancid on this box, following the instructions from this guide: http://www.linuxhomenetworking.com/wiki/index.php/Quick_HOWTO_:_Ch1_:_Network_Backups_With_Rancid All was going well up until I got to the part where I'm supposed to run the rancid-cvs command. At first I got the error 'cvs: command not found' and quickly realised that I needed to install something else, so a quick 'yum install cvs' fixed that. However, now when I run the rancid-cvs command, nothing happens, no error, no output at all, certainly not the output shown in the guide there. I tried re-compiling/installing rancid from the beginning in case something got missed due to cvs not being present earlier, but same problem. Any idea on what I can do to try and diagnose the problem? I'm running Fedora 10. to make sure nothing was happening instead of it working, just no output, I checked to see if the /usr/local/rancid//var/CVS/networking/ directory was created, it wasn't.
|
# ? Mar 17, 2009 04:04 |
|
|
# ? May 10, 2024 03:21 |
|
A simple question that has been bugging me for a while, how do i search recursively with wildcards(*) using ls? I am trying to do the bash counterpart of C:\dir /s *.junk Have tried to read the documentation and looked around with google, but I am unable to find a simple answer for this.
|
# ? Mar 17, 2009 06:06 |