Re: need wget help: dos not resolt file name
Quoting. wget http://www.overlooksoft.com/packages/download?plat=lx64ext=rpm; On 28/07/2014 4:08 PM, ToddAndMargo wrote: Hi All, The nice folks at OverLookSoft set this file up so that it will download automatically with wget http://www.overlooksoft.com/packages/download?plat=lx64ext=rpm It get a file called: download?plat=lx64ext=rpm Firefox does resolve this. What am I doing wrong with wget? Many thanks, -T -- Steven Haigh Email: net...@crc.id.au Web: http://www.crc.id.au Phone: (03) 9001 6090 - 0412 935 897 Fax: (03) 8338 0299 signature.asc Description: OpenPGP digital signature
RE: TFTP boot
That should be:
hardware ethernet 08:00:2b:4c:a3:82;
-Original message-
From:Patrick Mahan ma...@mahan.org
Sent: Monday 28th July 2014 15:00
To: Mahmood Naderan nt_mahm...@yahoo.com
Cc: scientific-linux-users@fnal.gov
Subject: Re: TFTP boot
On Jul 27, 2014, at 1:01 PM, Mahmood Naderan nt_mahm...@yahoo.com
mailto:nt_mahm...@yahoo.com wrote:
Hello,
I want to setup a diskless node and it should boot from a server. The guide
in http://wiki.centos.org/HowTos/PXE/PXE_Setup
http://wiki.centos.org/HowTos/PXE/PXE_Setup explains DHCP setup. However I
want to assign static IP to the node. As a result, I have to add the node's
MAC address somewhere in the server's configuration file.
Where should I do that?
Try googling static dhcp lease linux.
For example on a RH/CentOS/Fedora platform acting as your DHCP server, in
/etc/dhcpd.conf:
#
# assign a static lease to our printer
#
host lj4100 {
hardware 08:00:2b:4c:a3:82;
fixed-address 192.168.1.125;
}
Patrick
Coming to you from deep inside Fortress Mahan
Re: udev persistent net rules erratic inconsistencies
On Mon, Jul 28, 2014 at 5:35 AM, SCHAER Frederic frederic.sch...@cea.fr wrote: Hi, From time to time, we get reboot issues with some machines, and each time it looks like there are duplicated persistent rules for the Ethernet devices : On which Scientific Linux release? And is it hardware, or a virtualized system with added network configuration tools like VMware-tools ? The second set of rules seem to overwrite the first one, and then we get issues with our network config. This does not happen on all nodes, just apparently to some random ones, sometimes. Do try and figure out which nodes those are, if there is common software on them. I do warn you that letting NetworkManager anywhere near a stable server configuration is like letting toddlers into a machine shop. It *will* break things.
RE: udev persistent net rules erratic inconsistencies
Hi, For this specific server, this is SL 6.2. This is a hardware machine, dell C6220. We don't have network-manager installed, and we don't want it : we do a minimal install and then install what we need. We know how bad NM can be ;) All the nodes have the same software installed on them except maybe the OS which might be SL6.5 or 6.2, but we only have DELL servers which handle the emX scheme, so at least I can say it's happening on DELLs ;) I haven't had enough cases to be able to find out anything common, but in the present case, the udev file was modified on may 5. 2014 and for instance we had no yum update around that date, which I thought could be a perfect trigger for udev reconfig. -Message d'origine- De : Nico Kadel-Garcia [mailto:nka...@gmail.com] Envoyé : lundi 28 juillet 2014 13:14 À : SCHAER Frederic Cc : scientific-linux-users@fnal.gov Objet : Re: udev persistent net rules erratic inconsistencies On Mon, Jul 28, 2014 at 5:35 AM, SCHAER Frederic frederic.sch...@cea.fr wrote: Hi, From time to time, we get reboot issues with some machines, and each time it looks like there are duplicated persistent rules for the Ethernet devices : On which Scientific Linux release? And is it hardware, or a virtualized system with added network configuration tools like VMware-tools ? The second set of rules seem to overwrite the first one, and then we get issues with our network config. This does not happen on all nodes, just apparently to some random ones, sometimes. Do try and figure out which nodes those are, if there is common software on them. I do warn you that letting NetworkManager anywhere near a stable server configuration is like letting toddlers into a machine shop. It *will* break things.
Re: TFTP boot
Thanks. It seems that DNSmsq is better than DHCP configuration.
Regards,
Mahmood
On Monday, July 28, 2014 12:25 PM, Bill Maidment b...@maidment.com.au wrote:
That should be:
hardware ethernet 08:00:2b:4c:a3:82;
-Original message-
From:Patrick Mahan ma...@mahan.org
Sent: Monday 28th July 2014 15:00
To: Mahmood Naderan nt_mahm...@yahoo.com
Cc: scientific-linux-users@fnal.gov
Subject: Re: TFTP boot
On Jul 27, 2014, at 1:01 PM, Mahmood Naderan nt_mahm...@yahoo.com
mailto:nt_mahm...@yahoo.com wrote:
Hello,
I want to setup a diskless node and it should boot from a server. The guide
in http://wiki.centos.org/HowTos/PXE/PXE_Setup
http://wiki.centos.org/HowTos/PXE/PXE_Setup explains DHCP setup. However I
want to assign static IP to the node. As a result, I have to add the node's
MAC address somewhere in the server's configuration file.
Where should I do that?
Try googling static dhcp lease linux.
For example on a RH/CentOS/Fedora platform acting as your DHCP server, in
/etc/dhcpd.conf:
#
# assign a static lease to our printer
#
host lj4100 {
hardware 08:00:2b:4c:a3:82;
fixed-address 192.168.1.125;
}
Patrick
Coming to you from deep inside Fortress Mahan
Network aliases
I have an apple os script, that creates an lo alias ( for example 127.112.113) and then ssh -p 127.112.113 rest of the command To begin with ifconfig doesn't work on creating aliases and according to rhel docs I need to copy ethXxx files if I want to create these aliases. Is there a better way, than to copy files, to achieve the same?
Re: initrd.img vmlinuz
In any case, that's a CentOS repository, not Scientific Linux. In any case, look at the reference distribution at http://ftp1.scientificlinux.org/linux/scientific/6.3/x86_64/os/Packages/ Excuse me, I don't see such packages in the link Regards, Mahmood On , Mahmood Naderan nt_mahm...@yahoo.com wrote: I need them for PXE boot. I prefer to have the same versions on nodes. Regards, Mahmood On Monday, July 28, 2014 3:56 PM, Nico Kadel-Garcia nka...@gmail.com wrote: On Sun, Jul 27, 2014 at 5:36 PM, Mahmood Naderan nt_mahm...@yahoo.com wrote: Hello, Where can I download the initrd.img and vmlinuz files for SL-6.3? Those files are only available for SL-6.5 http://mirror.de.leaseweb.net/centos/ Why would you *want* them? You'll miss out on all the patches since SL 6.3?! In any case, that's a CentOS repository, not Scientific Linux. In any case, look at the reference distribution at http://ftp1.scientificlinux.org/linux/scientific/6.3/x86_64/os/Packages/
Re: initrd.img vmlinuz
Hi, you can find them here: http://ftp1.scientificlinux.org/linux/scientific/6.3/x86_64/os/images/pxeboot/ Regards, Jan. On 28.07.2014, at 16:09, Mahmood Naderan wrote: In any case, that's a CentOS repository, not Scientific Linux. In any case, look at the reference distribution at http://ftp1.scientificlinux.org/linux/scientific/6.3/x86_64/os/Packages/ Excuse me, I don't see such packages in the link Regards, Mahmood On , Mahmood Naderan nt_mahm...@yahoo.com wrote: I need them for PXE boot. I prefer to have the same versions on nodes. Regards, Mahmood On Monday, July 28, 2014 3:56 PM, Nico Kadel-Garcia nka...@gmail.com wrote: On Sun, Jul 27, 2014 at 5:36 PM, Mahmood Naderan nt_mahm...@yahoo.com wrote: Hello, Where can I download the initrd.img and vmlinuz files for SL-6.3? Those files are only available for SL-6.5 http://mirror.de.leaseweb.net/centos/ Why would you *want* them? You'll miss out on all the patches since SL 6.3?! In any case, that's a CentOS repository, not Scientific Linux. In any case, look at the reference distribution at http://ftp1.scientificlinux.org/linux/scientific/6.3/x86_64/os/Packages/
Re: Network aliases
On 07/28/2014 08:08 AM, Andrew Z wrote: I have an apple os script, that creates an lo alias ( for example 127.112.113) and then ssh -p 127.112.113 rest of the command To begin with ifconfig doesn't work on creating aliases and according to rhel docs I need to copy ethXxx files if I want to create these aliases. Is there a better way, than to copy files, to achieve the same? What's the concern with making a new ifcfg-* file to create the alias? -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA, Boulder/CoRA Office FAX: 303-415-9702 3380 Mitchell Lane or...@nwra.com Boulder, CO 80301 http://www.nwra.com
Re: XFS and dump?
Hi ToddAndMargo! On 2014.07.27 at 22:25:08 -0700, ToddAndMargo wrote next: Why would you *want* that? Seriously? When backups of files, rather than filesystems, are much more easily mirrored onto a read-only, accessible target to allow people to recover their deleted files quickly and cheaply? Even if you need to back up to tape, AMANDA and half a dozen other technologies do a much better job with tar or rsync based backups. Hi Nico, 1) It is *insanely* fast 2) it is command line 3) I am the one that does the recovering, not the user. (I leave the backup drives unmounted when not on use on purpose! I do not want the user anywhere near those drives.) Dump was a stupid program in the first place. Leave it behind. Dump may work fine for you a thousand times. But it _will_ fail under the right circumstances. And there is nothing you can do about it. http://lwn.net/2001/0503/a/lt-dump.php3 Seriously, don't use dump. Backup software that highly depends on fs implementation is a very bad idea in modern world. Accessing data at low level from fs mounted in r/w (i.e. the data might be modified, so special tricks are required to extract meaningful copy) is just asking for trouble. Note also that dump is 1) highly unportable (in Linux: ext2/3/4 only, *BSD dump isn't compatible with Linux version and so on) 2) not supported in other UNIX operating systems anymore because its design doesn't work with modern file system design. dump/restore don't exist in Solaris 11 at all (official recommendation: zfs dump/restore for whole backups, cpio/tar/rdiff-backup/areca/etc for more agile backups). For AIX IBM recommends tar and cpio for file-level backup as modern replacements (http://www.ibmsystemsmag.com/aix/administrator/backuprecovery/Backup-and-Restore-With-AIX) If are 100% sure that you need fs-level backup instead of file-level backup, then don't ask for things impossible by definition (restoring fs-level backup to another fs). Yes, dump/restore are fast but they are able to do it by duplicating parts of kernel fs implementation in userspace, which is why there will be irresistible problems with them. -- Vladimir
Re: initrd.img vmlinuz
Thank you very much Regards, Mahmood On Monday, July 28, 2014 7:07 PM, Jan A. Stillings stilli...@physik.uni-bonn.de wrote: Hi, you can find them here: http://ftp1.scientificlinux.org/linux/scientific/6.3/x86_64/os/images/pxeboot/ Regards, Jan. On 28.07.2014, at 16:09, Mahmood Naderan wrote: In any case, that's a CentOS repository, not Scientific Linux. In any case, look at the reference distribution at http://ftp1.scientificlinux.org/linux/scientific/6.3/x86_64/os/Packages/ Excuse me, I don't see such packages in the link Regards, Mahmood On , Mahmood Naderan nt_mahm...@yahoo.com wrote: I need them for PXE boot. I prefer to have the same versions on nodes. Regards, Mahmood On Monday, July 28, 2014 3:56 PM, Nico Kadel-Garcia nka...@gmail.com wrote: On Sun, Jul 27, 2014 at 5:36 PM, Mahmood Naderan nt_mahm...@yahoo.com wrote: Hello, Where can I download the initrd.img and vmlinuz files for SL-6.3? Those files are only available for SL-6.5 http://mirror.de.leaseweb.net/centos/ Why would you *want* them? You'll miss out on all the patches since SL 6.3?! In any case, that's a CentOS repository, not Scientific Linux. In any case, look at the reference distribution at http://ftp1.scientificlinux.org/linux/scientific/6.3/x86_64/os/Packages/
Re: need wget help: dos not resolt file name
On 28/07/2014 4:08 PM, ToddAndMargo wrote: Hi All, The nice folks at OverLookSoft set this file up so that it will download automatically with wget http://www.overlooksoft.com/packages/download?plat=lx64ext=rpm It get a file called: download?plat=lx64ext=rpm Firefox does resolve this. What am I doing wrong with wget? Many thanks, -T On 07/27/2014 11:12 PM, Steven Haigh wrote: Quoting. wget http://www.overlooksoft.com/packages/download?plat=lx64ext=rpm; Hi Steven, Firefox sees the same thing, but resolves it to a real file name. I am trying to duplicate that with wget. -T
Re: XFS and dump?
On 07/28/2014 09:12 AM, Vladimir Mosgalin wrote: Hi ToddAndMargo! On 2014.07.27 at 22:25:08 -0700, ToddAndMargo wrote next: Why would you *want* that? Seriously? When backups of files, rather than filesystems, are much more easily mirrored onto a read-only, accessible target to allow people to recover their deleted files quickly and cheaply? Even if you need to back up to tape, AMANDA and half a dozen other technologies do a much better job with tar or rsync based backups. Hi Nico, 1) It is *insanely* fast 2) it is command line 3) I am the one that does the recovering, not the user. (I leave the backup drives unmounted when not on use on purpose! I do not want the user anywhere near those drives.) Dump was a stupid program in the first place. Leave it behind. Dump may work fine for you a thousand times. But it _will_ fail under the right circumstances. And there is nothing you can do about it. http://lwn.net/2001/0503/a/lt-dump.php3 Seriously, don't use dump. Backup software that highly depends on fs implementation is a very bad idea in modern world. Accessing data at low level from fs mounted in r/w (i.e. the data might be modified, so special tricks are required to extract meaningful copy) is just asking for trouble. Note also that dump is 1) highly unportable (in Linux: ext2/3/4 only, *BSD dump isn't compatible with Linux version and so on) 2) not supported in other UNIX operating systems anymore because its design doesn't work with modern file system design. dump/restore don't exist in Solaris 11 at all (official recommendation: zfs dump/restore for whole backups, cpio/tar/rdiff-backup/areca/etc for more agile backups). For AIX IBM recommends tar and cpio for file-level backup as modern replacements (http://www.ibmsystemsmag.com/aix/administrator/backuprecovery/Backup-and-Restore-With-AIX) If are 100% sure that you need fs-level backup instead of file-level backup, then don't ask for things impossible by definition (restoring fs-level backup to another fs). Yes, dump/restore are fast but they are able to do it by duplicating parts of kernel fs implementation in userspace, which is why there will be irresistible problems with them. Hi Vladimir, Is xfsdump any better? What do you prefer in its place? It looks like a full wipe to go to 7 and xfs anyway. -T
Re: need wget help: dos not resolt file name
On 29/07/2014 2:54 AM, ToddAndMargo wrote: On 28/07/2014 4:08 PM, ToddAndMargo wrote: Hi All, The nice folks at OverLookSoft set this file up so that it will download automatically with wget http://www.overlooksoft.com/packages/download?plat=lx64ext=rpm It get a file called: download?plat=lx64ext=rpm Firefox does resolve this. What am I doing wrong with wget? Many thanks, -T On 07/27/2014 11:12 PM, Steven Haigh wrote: Quoting. wget http://www.overlooksoft.com/packages/download?plat=lx64ext=rpm; Hi Steven, Firefox sees the same thing, but resolves it to a real file name. I am trying to duplicate that with wget. Yeah, wget isn't that smart... I've never figured it out - When I know the filename, I just use the -O filename option to override the output file... -- Steven Haigh Email: net...@crc.id.au Web: http://www.crc.id.au Phone: (03) 9001 6090 - 0412 935 897 Fax: (03) 8338 0299 signature.asc Description: OpenPGP digital signature
Re: initrd.img vmlinuz
On Mon, Jul 28, 2014 at 12:46 PM, Mahmood Naderan nt_mahm...@yahoo.com wrote: Thank you very much Regards, Mahmood For reference, they are also always inside 'boot.iso' image and the installation CD or DVD's.
Re: need wget help: dos not resolt file name
Hi, On 2014-07-28 2:01 PM, Steven Haigh wrote: On 29/07/2014 2:54 AM, ToddAndMargo wrote: On 28/07/2014 4:08 PM, ToddAndMargo wrote: Hi All, The nice folks at OverLookSoft set this file up so that it will download automatically with wget http://www.overlooksoft.com/packages/download?plat=lx64ext=rpm It get a file called: download?plat=lx64ext=rpm Firefox does resolve this. What am I doing wrong with wget? Many thanks, -T On 07/27/2014 11:12 PM, Steven Haigh wrote: Quoting. wget http://www.overlooksoft.com/packages/download?plat=lx64ext=rpm; Hi Steven, Firefox sees the same thing, but resolves it to a real file name. I am trying to duplicate that with wget. Yeah, wget isn't that smart... I've never figured it out - When I know the filename, I just use the -O filename option to override the output file... Try: wget --content-disposition http://www.overlooksoft.com/packages/download?plat=lx64ext=rpm; The --content-disposition flag causes wget to use the filename provided by the Content-Disposition HTTP header, if any. It's experimental and apparently doesn't always work, but in this case, it worked for me. The man page for wget says: --content-disposition If this is set to on, experimental (not fully-functional) support for Content-Disposition headers is enabled. This can currently result in extra round-trips to the server for a HEAD request, and is known to suffer from a few bugs, which is why it is not currently enabled by default. This option is useful for some file-downloading CGI programs that use Content-Disposition headers to describe what the name of a downloaded file should be. Cheers, John Roberts SNOLAB
Re: TFTP boot
On Sun, Jul 27, 2014 at 01:01:30PM -0700, Mahmood Naderan wrote: Hello, I want to setup a diskless node and it should boot from a server. The guide in http://wiki.centos.org/HowTos/PXE/PXE_Setup explains DHCP setup. However I want to assign static IP to the node. As a result, I have to add the node's MAC address somewhere in the server's configuration file. Where should I do that? Please look at the documentation here - provided are the snippets for dhcpd.conf, pxelinux config files, etc: http://www.triumf.info/wiki/DAQwiki/index.php/VME-CPU#Network_boot -- Konstantin Olchanski Data Acquisition Systems: The Bytes Must Flow! Email: olchansk-at-triumf-dot-ca Snail mail: 4004 Wesbrook Mall, TRIUMF, Vancouver, B.C., V6T 2A3, Canada
Re: udev persistent net rules erratic inconsistencies
Hi, there - persistent-net.rules is a clever scheme to prevent your computer
from booting after you replace a burned out motherboard. To disable it, run:
touch /etc/udev/rules.d/75-persistent-net-generator.rules
rm /etc/udev/rules.d/70-persistent-net.rules
K.O.
On Mon, Jul 28, 2014 at 09:35:17AM +, SCHAER Frederic wrote:
Hi,
From time to time, we get reboot issues with some machines, and each time it
looks like there are duplicated persistent rules for the Ethernet devices :
cat /etc/udev/rules.d/70-persistent-net.rules
# This file was automatically generated by the /lib/udev/write_net_rules
# program, run by the persistent-net-generator.rules rules file.
#
# You can modify it, as long as you keep each rule on a single
# line, and change only the value of the NAME= key.
# PCI device 0x8086:0x1521 (igb) (custom name provided by external tool)
SUBSYSTEM==net, ACTION==add, DRIVERS==?*,
ATTR{address}==84:8f:69:fb:c1:2a, ATTR{type}==1, KERNEL==eth*,
NAME=em1
# PCI device 0x8086:0x1521 (igb) (custom name provided by external tool)
SUBSYSTEM==net, ACTION==add, DRIVERS==?*,
ATTR{address}==84:8f:69:fb:c1:2b, ATTR{type}==1, KERNEL==eth*,
NAME=em2
# PCI device 0x8086:0x1521 (igb) (custom name provided by external tool)
SUBSYSTEM==net, ACTION==add, DRIVERS==?*,
ATTR{address}==84:8f:69:fb:c1:2a, ATTR{type}==1, KERNEL==eth*,
NAME=eth0
# PCI device 0x8086:0x1521 (igb) (custom name provided by external tool)
SUBSYSTEM==net, ACTION==add, DRIVERS==?*,
ATTR{address}==84:8f:69:fb:c1:2b, ATTR{type}==1, KERNEL==eth*,
NAME=eth1
The second set of rules seem to overwrite the first one, and then we get
issues with our network config.
This does not happen on all nodes, just apparently to some random ones,
sometimes.
I'm wondering if some of you might have faced and solved that erratic thing
already ?
We want to keep the emX scheme for nodes which support it...
Thanks
--
Konstantin Olchanski
Data Acquisition Systems: The Bytes Must Flow!
Email: olchansk-at-triumf-dot-ca
Snail mail: 4004 Wesbrook Mall, TRIUMF, Vancouver, B.C., V6T 2A3, Canada
Re: Network aliases
I havent figured out how to manage those files.. The original script creates dozens of aliases - one per destination, by replacing first Xxx with 127. This makes it intuitive where alias points to. Now with each file having the same name (lo:port) imnot sure how to manage these aliases. I want to add, that I suspect that im missing something obvious and/or just dont understand how network stack works now. Any help is very much appreciated, since all my information is from rhel one page guide , some blog and just trying to get aliases to work. On Jul 28, 2014 10:40 AM, Orion Poplawski or...@cora.nwra.com wrote: On 07/28/2014 08:08 AM, Andrew Z wrote: I have an apple os script, that creates an lo alias ( for example 127.112.113) and then ssh -p 127.112.113 rest of the command To begin with ifconfig doesn't work on creating aliases and according to rhel docs I need to copy ethXxx files if I want to create these aliases. Is there a better way, than to copy files, to achieve the same? What's the concern with making a new ifcfg-* file to create the alias? -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA, Boulder/CoRA Office FAX: 303-415-9702 3380 Mitchell Lane or...@nwra.com Boulder, CO 80301 http://www.nwra.com
Re: Network aliases
On 2014-07-28 16:08, Andrew Z wrote: I have an apple os script, that creates an lo alias ( for example 127.112.113) and then ssh -p 127.112.113 rest of the command To begin with ifconfig doesn't work on creating aliases and according to rhel docs I need to copy ethXxx files if I want to create these aliases. Is there a better way, than to copy files, to achieve the same? It's as simple as on apple/bsd. ifconfig ethX:Y inet $IP/$CIDR ifconfig ethX:Y del $IP -- olli
Re: Network aliases
For the exception that ifconfig is no longer there.. On Jul 28, 2014 4:55 PM, olli hauer oha...@gmx.de wrote: On 2014-07-28 16:08, Andrew Z wrote: I have an apple os script, that creates an lo alias ( for example 127.112.113) and then ssh -p 127.112.113 rest of the command To begin with ifconfig doesn't work on creating aliases and according to rhel docs I need to copy ethXxx files if I want to create these aliases. Is there a better way, than to copy files, to achieve the same? It's as simple as on apple/bsd. ifconfig ethX:Y inet $IP/$CIDR ifconfig ethX:Y del $IP -- olli
Re: Network aliases
On 07/28/2014 01:49 PM, Andrew Z wrote: I havent figured out how to manage those files.. The original script creates dozens of aliases - one per destination, by replacing first Xxx with 127. This makes it intuitive where alias points to. Now with each file having the same name (lo:port) imnot sure how to manage these aliases. I want to add, that I suspect that im missing something obvious and/or just dont understand how network stack works now. Any help is very much appreciated, since all my information is from rhel one page guide , some blog and just trying to get aliases to work. Manually you could do: ip addr add 127.112.113.xxx dev lo On Jul 28, 2014 10:40 AM, Orion Poplawski or...@cora.nwra.com mailto:or...@cora.nwra.com wrote: On 07/28/2014 08:08 AM, Andrew Z wrote: I have an apple os script, that creates an lo alias ( for example 127.112.113) and then ssh -p 127.112.113 rest of the command To begin with ifconfig doesn't work on creating aliases and according to rhel docs I need to copy ethXxx files if I want to create these aliases. Is there a better way, than to copy files, to achieve the same? What's the concern with making a new ifcfg-* file to create the alias? -- Orion Poplawski Technical Manager 303-415-9701 x222 tel:303-415-9701%20x222 NWRA, Boulder/CoRA Office FAX: 303-415-9702 tel:303-415-9702 3380 Mitchell Lane or...@nwra.com mailto:or...@nwra.com Boulder, CO 80301 http://www.nwra.com -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA, Boulder/CoRA Office FAX: 303-415-9702 3380 Mitchell Lane or...@nwra.com Boulder, CO 80301 http://www.nwra.com
Re: udev persistent net rules erratic inconsistencies
On Mon, Jul 28, 2014 at 5:35 AM, SCHAER Frederic frederic.sch...@cea.fr wrote:
From time to time, we get reboot issues with some machines, and each time it
looks like there are duplicated persistent rules for the Ethernet devices :
cat /etc/udev/rules.d/70-persistent-net.rules
# This file was automatically generated by the /lib/udev/write_net_rules
# program, run by the persistent-net-generator.rules rules file.
# You can modify it, as long as you keep each rule on a single
# line, and change only the value of the NAME= key.
# PCI device 0x8086:0x1521 (igb) (custom name provided by external tool)
SUBSYSTEM==net, ACTION==add, DRIVERS==?*,
ATTR{address}==84:8f:69:fb:c1:2a, ATTR{type}==1, KERNEL==eth*,
NAME=em1
# PCI device 0x8086:0x1521 (igb) (custom name provided by external tool)
SUBSYSTEM==net, ACTION==add, DRIVERS==?*,
ATTR{address}==84:8f:69:fb:c1:2b, ATTR{type}==1, KERNEL==eth*,
NAME=em2
# PCI device 0x8086:0x1521 (igb) (custom name provided by external tool)
SUBSYSTEM==net, ACTION==add, DRIVERS==?*,
ATTR{address}==84:8f:69:fb:c1:2a, ATTR{type}==1, KERNEL==eth*,
NAME=eth0
# PCI device 0x8086:0x1521 (igb) (custom name provided by external tool)
SUBSYSTEM==net, ACTION==add, DRIVERS==?*,
ATTR{address}==84:8f:69:fb:c1:2b, ATTR{type}==1, KERNEL==eth*,
NAME=eth1
The second set of rules seem to overwrite the first one, and then we get
issues with our network config.
This does not happen on all nodes, just apparently to some random ones,
sometimes.
I’m wondering if some of you might have faced and solved that erratic thing
already ?
We want to keep the emX scheme for nodes which support it…
The emX names come from biosdevname (part of a base install):
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Deployment_Guide/appe-Consistent_Network_Device_Naming.html
You should be able to delete the last two rules, the ethX rules, for
your interfaces to use the emX names, unless you're using
biosdevname=0 on the kernel cmdline.
