Re: [arch-general] Last networkmanager in testing busts everything in?gnome ?
On Wed, Jan 20, 2010 at 07:53:10AM -0500, Baho Utot wrote: On Friday 08 January 2010 03:54:23 Allan McRae wrote: Ionut Biru wrote: Yep, my /etc/hosts file was definitely changed: # # /etc/hosts: static lookup table for host names # 127.0.0.1 localhost.localdomain localhost #ip-address hostname.domain.org hostname 127.0.0.1 arch 127.0.0.1 localhost.localdomain localhost I was having trouble with networking several years ago and had a /etc/host like yours. I posted to usenet and the network gurus there promptly busted me for that layout. They told me to do this instead: #ip-address hostname.domain.org hostname 127.0.0.1 localhost.localdomainlocalhost 127.0.0.2 arch.yourdomain arch They claim that the above conforms to the RFC's and reusing the 127.0.0.1 address can confuse some apps. I have been using the above and it has always worked, no busted apps. This also seems to work for me, at least the extra aliases at the end of the line are respected: #ip-address hostname.domain.org hostname 127.0.0.1 localhost.localdomain localhostarch -- Jim Pryor prof...@jimpryor.net
Re: [arch-general] Can you rebuild an nv dmraid array in linux - or is it just gparted and copy partitions?
David C. Rankin wrote: Listmates I had a failing drive in my arch server so I installed a new one. The bios raid 'nvidia dmraid' provides the feature to 'Rebuild' the array and allows for adding the new disk to the array. After that configuration you are prompted to Boot to an OS that supports rebuilding.. (or something very close to that). I image that there is some windows utility that does the rebuild on boot. Something probably on one of the cd's that came with the motherboard. The motherboard is an MSI K9N2 SLI Platinum (MS-7374) and has the nvidia dmraid 0/1/5/ chipset. In the past I have always just used gparted to copy partitions from good drive - new drive and created a new array. Based on the output of creating the array in the bios, I'm curious if I can just rebuild does it work in linux? how? This is not an answer to your question directly... I have hardware raid on my desktop but I use software raid, raid5 and raid 1 arrays using mdadm. The reason I don't use the on board raid or a hardware raid is that the different manufactures use different methods causing problems when you want move the array to another system or the raid card fails. Then you usally lose the entire raid array. With software raid you can move the raid to another system and it uses whatever stat hardware that is in the system.. ie lose the stata card just install another one and you are good. To replace a failed drive with software raid all that needs to be done is to fail the drive using mdadm then physical remove and replace the drive. Then use sfdisk to copy the partition setup from one of the drives in the raid to the new drive. After the partitioning is complete add the drive to the array using mdadm and your done. All that you need to do is to monitor the rebuilding by a simple cat /proc/mdstat.
Re: [arch-general] Last networkmanager in testing busts everything in?gnome ?
Jim Pryor wrote: On Wed, Jan 20, 2010 at 07:53:10AM -0500, Baho Utot wrote: On Friday 08 January 2010 03:54:23 Allan McRae wrote: Ionut Biru wrote: Yep, my /etc/hosts file was definitely changed: # # /etc/hosts: static lookup table for host names # 127.0.0.1 localhost.localdomain localhost #ip-address hostname.domain.org hostname 127.0.0.1 arch 127.0.0.1 localhost.localdomain localhost I was having trouble with networking several years ago and had a /etc/host like yours. I posted to usenet and the network gurus there promptly busted me for that layout. They told me to do this instead: #ip-address hostname.domain.org hostname 127.0.0.1 localhost.localdomainlocalhost 127.0.0.2 arch.yourdomain arch They claim that the above conforms to the RFC's and reusing the 127.0.0.1 address can confuse some apps. I have been using the above and it has always worked, no busted apps. This also seems to work for me, at least the extra aliases at the end of the line are respected: #ip-address hostname.domain.org hostname 127.0.0.1 localhost.localdomain localhostarch Yes taht is a good way also but I like to keep all the aliases on there own line, that way I can assign a domain to the host.
Re: [arch-general] [signoff] kernel 2.6.32.4-1 [ Possible Problems... ]
After further testing with kernel 2.6.32.4-1, I have found two bugs: (1) the kernel upgrade kills WindowMaker (2) the kernel upgrade kills VirtualBox even with 2.6.32.3, I had to recomile virtualbox module. I tried to start a VM and system froze solid and had to do hard reset. With a recompile, its working though. Are you talking about recompile of the vbox module, that I just did, or are you talking about a kernel recompile? vbox module recompile. It fixed the issue on 2.6.32-3 but from your post it seems, it is not sufficient for 2.6.32-4. It is working here with x86_64. I have updated _only_ the kernel, headers and firmware, after that I've rebuilt the vbox module and reboot. Everything is working as it should. It is clearly stated when you install virtualbox that you need to rebuild the vbox module after every kernel update, if before it was only needed after a kernel version change it was just luck. And I guess just rebuilding the module doesn't cut it, if it is already inserted you may have to rmmod it and modprobe it again, but to be safe I think it is better to reboot the machine.
[arch-general] Arch Linux Press Review wiki page updated
People, I've updated the wiki page [1] of the press review with the 2010 section and the interview to some people of the dev team of Arch. If you have blogs, sites, etc where Arch was mentioned and it's not showing on that list, you can edit the wiki, and contribute with the wiki!. I've to confess that I love to read those reviews :). Cheers! and nice interview btw [1] http://wiki.archlinux.org/index.php/Arch_Linux_Press_Review -- Angel Velásquez angvp @ irc.freenode.net Arch Linux Trusted User Linux Counter: #359909 http://www.angvp.com
Re: [arch-general] Can you rebuild an nv dmraid array in linux - or is it just gparted and copy partitions?
On 01/21/2010 05:38 AM, Baho Utot wrote: This is not an answer to your question directly... I have hardware raid on my desktop but I use software raid, raid5 and raid 1 arrays using mdadm. The reason I don't use the on board raid or a hardware raid is that the different manufactures use different methods causing problems when you want move the array to another system or the raid card fails. Then you usally lose the entire raid array. With software raid you can move the raid to another system and it uses whatever stat hardware that is in the system.. ie lose the stata card just install another one and you are good. To replace a failed drive with software raid all that needs to be done is to fail the drive using mdadm then physical remove and replace the drive. Then use sfdisk to copy the partition setup from one of the drives in the raid to the new drive. After the partitioning is complete add the drive to the array using mdadm and your done. All that you need to do is to monitor the rebuilding by a simple cat /proc/mdstat. Baho, Thanks. I am arriving at that conclusion as well. Currently I have 6 onboard raid sets and 4 software sets. I have had good luck replacing failed disks with both in the past. I agree mdraid is much more flexible than dmraid and from the testing I've done between fake-raid and software raid there is 0 (zero) performance difference between the two. I'll let you know how this replacement goes. I'm in the process or copying partitions with gparted as we speak... -- David C. Rankin, J.D.,P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com
Re: [arch-general] [signoff] kernel 2.6.32.4-1 [ Possible Problems... ]
On 01/20/2010 01:44 PM, Ionut Biru wrote: packages for e17 wasn't rebuild. have patience or downgrade to extra/community and don't flood the email with useless problems. The rebuilding process is NOT FINISHED. Sorry, those weren't complaints... Since I have blackbox, e16, e17, fluxbox, fvwm2 gnome, (the old) kdemod3, kde4, openbox, sawfish, xfce, and WindowMaker on the box, I was just checking to see which worked after the kernel update and providing feeback. Note to self: Premise responses following new kernel install with These just may not have been rebuilt yet but ... :p -- David C. Rankin, J.D.,P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com
Re: [arch-general] [signoff] kernel 2.6.32.4-1 [ Possible Problems... ]
On Thu, Jan 21, 2010 at 01:19:44PM -0600, David C. Rankin wrote: On 01/20/2010 01:44 PM, Ionut Biru wrote: packages for e17 wasn't rebuild. have patience or downgrade to extra/community and don't flood the email with useless problems. The rebuilding process is NOT FINISHED. Sorry, those weren't complaints... Since I have blackbox, e16, e17, fluxbox, fvwm2 gnome, (the old) kdemod3, kde4, openbox, sawfish, xfce, and WindowMaker on the box, I was just checking to see which worked after the kernel update and providing feeback. I'm guessing that none of those were broken because of the kernel update, but because of the unfinished libpng rebuild in testing. -- Byron Clark
[arch-general] unclean filesystem
Uh, this might be a quick question, since it's a little abstract. I'm dual booting and I've noticed that every time I boot into Arch after I've booted into one of my other systems, there is a forced file system check. It's not a huge deal, because I use Arch almost exclusively, but I have to say, I cringe at the thought of having to boot into one of the other systems now. How would you trouble shoot this? I have also never seen a system where after every ~20 reboots, a file system check is mandatory. This might just be my ignorance here, so a reference to some background about that part of the boot process might suffice. Thanks in advance. -- chris
Re: [arch-general] unclean filesystem
On 01/21/2010 09:11 PM, christopher floess wrote: Uh, this might be a quick question, since it's a little abstract. I'm dual booting and I've noticed that every time I boot into Arch after I've booted into one of my other systems, there is a forced file system check. It's not a huge deal, because I use Arch almost exclusively, but I have to say, I cringe at the thought of having to boot into one of the other systems now. How would you trouble shoot this? I have also never seen a system where after every ~20 reboots, a file system check is mandatory. This might just be my ignorance here, so a reference to some background about that part of the boot process might suffice. Thanks in advance. -- chris Is the other operative system a windows? Maybe the hardware clock is continuously changing, so the last write looks like in the future? -- Enrico Carlessoecarle...@ecarlesso.org http://www.ecarlesso.org carlessoenrico @ skype enricocarle...@gmail.com @ gtalk msn
Re: [arch-general] [signoff] kernel 2.6.32.4-1 [ Possible Problems... ]
On 01/21/2010 01:31 PM, Byron Clark wrote: I'm guessing that none of those were broken because of the kernel update, but because of the unfinished libpng rebuild in testing. 100% Correct. The issues were libpng12 to libpng14 and libjpg. -- David C. Rankin, J.D.,P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com
Re: [arch-general] unclean filesystem
On Fri, Jan 22, 2010 at 4:45 AM, Enrico Carlesso enr...@ecarlesso.org wrote: On 01/21/2010 09:11 PM, christopher floess wrote: Uh, this might be a quick question, since it's a little abstract. I'm dual booting and I've noticed that every time I boot into Arch after I've booted into one of my other systems, there is a forced file system check. It's not a huge deal, because I use Arch almost exclusively, but I have to say, I cringe at the thought of having to boot into one of the other systems now. How would you trouble shoot this? I have also never seen a system where after every ~20 reboots, a file system check is mandatory. This might just be my ignorance here, so a reference to some background about that part of the boot process might suffice. Thanks in advance. -- chris Is the other operative system a windows? Maybe the hardware clock is continuously changing, so the last write looks like in the future? ^ what he said Also, the hard disk check can be set. I set mine to every 30 days (previously was rebooting up to 10 times a day from playing with kernel configs). Its not a hard-coded check, and can be (I think) cancelled with Ctrl-C, based on some other mails going around here a few days back.
Re: [arch-general] unclean filesystem
On 01/21/2010 01:11 PM, christopher floess wrote: Uh, this might be a quick question, since it's a little abstract. I'm dual booting and I've noticed that every time I boot into Arch after I've booted into one of my other systems, there is a forced file system check. It's not a huge deal, because I use Arch almost exclusively, but I have to say, I cringe at the thought of having to boot into one of the other systems now. How would you trouble shoot this? I have also never seen a system where after every ~20 reboots, a file system check is mandatory. This might just be my ignorance here, so a reference to some background about that part of the boot process might suffice. Thanks in advance. -- chris Are you using shut down or hibernate? I think if you hibernate Windows, your partitions don't get unmounted and that could be what's happening.