Re: [CentOS] the wandering MAC?
On Wed, Oct 13, 2010 at 12:28 PM, Boris Epstein borepst...@gmail.com wrote: Hi all, I have what appears to be a truly puzzling problem. I've got this P4 32-bit machine running CentOS 5.5 with XEN that has two NICs: one onboard, an Intel Corporation 82540EM Gigabit and one on an expansion card, Realtek Semiconductor Co., Ltd. RTL-8169 Gigabit. The second one is recognized as eth1. What's happening is, it is showing up under one of the two MAC's: either 00:0a:cd:1a:c1:71 or 00:00:00:00:c1:71. If you reboot it the MAC stays the same; if you shutdown and do a full powerdown it seems to change. Obviously, after it goes from one MAC to another you have to play with the start-up scripts for this interface to start up correctly and this becomes a major annoyance. Any idea what all of this mess could mean? Thanks. Boris. OK, people, here's something looks promising: http://www.centos.org/modules/newbb/viewtopic.php?viewmode=flatorder=ASCtopic_id=19571forum=40#forumpost73378 Looks like there is a whole special repo for this sort of drivers. Has anybody used it? How is it? Anyways, I think I'd give it a try. Cheers, Boris. ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On Mon, Nov 29, 2010 at 6:58 PM, Boris Epstein borepst...@gmail.com wrote: http://www.centos.org/modules/newbb/viewtopic.php?viewmode=flatorder=ASCtopic_id=19571forum=40#forumpost73378 Looks like there is a whole special repo for this sort of drivers. Has anybody used it? How is it? Elrepo is trustworthy. Got myself out of an realsh..err,tek problem using their packages, too. Go ahead. ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
[CentOS] the wandering MAC?
Hi all, I have what appears to be a truly puzzling problem. I've got this P4 32-bit machine running CentOS 5.5 with XEN that has two NICs: one onboard, an Intel Corporation 82540EM Gigabit and one on an expansion card, Realtek Semiconductor Co., Ltd. RTL-8169 Gigabit. The second one is recognized as eth1. What's happening is, it is showing up under one of the two MAC's: either 00:0a:cd:1a:c1:71 or 00:00:00:00:c1:71. If you reboot it the MAC stays the same; if you shutdown and do a full powerdown it seems to change. Obviously, after it goes from one MAC to another you have to play with the start-up scripts for this interface to start up correctly and this becomes a major annoyance. Any idea what all of this mess could mean? Thanks. Boris. ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On 10/13/2010 09:28 AM, Boris Epstein wrote: What's happening is, it is showing up under one of the two MAC's: either 00:0a:cd:1a:c1:71 or 00:00:00:00:c1:71. If you reboot it the MAC stays the same; if you shutdown and do a full powerdown it seems to change. I would say the card is probably dying and replace it. -- Benjamin Franz ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On Wed, Oct 13, 2010 at 12:35 PM, Benjamin Franz jfr...@freerun.com wrote: On 10/13/2010 09:28 AM, Boris Epstein wrote: What's happening is, it is showing up under one of the two MAC's: either 00:0a:cd:1a:c1:71 or 00:00:00:00:c1:71. If you reboot it the MAC stays the same; if you shutdown and do a full powerdown it seems to change. I would say the card is probably dying and replace it. -- Benjamin Franz ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos Thanks. That's possible, sure. I wonder though - it seems to work just fine when it's up, pretty fast, no abnormal error rate, it is brand new. But you could be right, of course. Boris. ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On 13/10/2010 18:37, Boris Epstein wrote: On Wed, Oct 13, 2010 at 12:35 PM, Benjamin Franzjfr...@freerun.com wrote: On 10/13/2010 09:28 AM, Boris Epstein wrote: What's happening is, it is showing up under one of the two MAC's: either 00:0a:cd:1a:c1:71 or 00:00:00:00:c1:71. If you reboot it the MAC stays the same; if you shutdown and do a full powerdown it seems to change. I would say the card is probably dying and replace it. Thanks. That's possible, sure. I wonder though - it seems to work just fine when it's up, pretty fast, no abnormal error rate, it is brand new. But you could be right, of course. I've tended to find that when a card is failing the MAC address starts setting itself to FF:FF:FF:FF:FF:FF not 00:00:00:XX:XX:XX The first three bytes are Vendor ID on a MAC address, you haven't got anything in there that might fiddle with that? Is it an OEM card? ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On 10/13/2010 06:46 PM, Giles Coochey wrote: On 13/10/2010 18:37, Boris Epstein wrote: On Wed, Oct 13, 2010 at 12:35 PM, Benjamin Franzjfr...@freerun.com wrote: On 10/13/2010 09:28 AM, Boris Epstein wrote: What's happening is, it is showing up under one of the two MAC's: either 00:0a:cd:1a:c1:71 or 00:00:00:00:c1:71. If you reboot it the MAC stays the same; if you shutdown and do a full powerdown it seems to change. I would say the card is probably dying and replace it. Thanks. That's possible, sure. I wonder though - it seems to work just fine when it's up, pretty fast, no abnormal error rate, it is brand new. But you could be right, of course. I've tended to find that when a card is failing the MAC address starts setting itself to FF:FF:FF:FF:FF:FF not 00:00:00:XX:XX:XX FF:FF:FF:FF:FF:FF is broadcast. The first three bytes are Vendor ID on a MAC address, you haven't got anything in there that might fiddle with that? Is it an OEM card? Timo ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On 13/10/2010 19:00, Timo Schoeler wrote: On 10/13/2010 06:46 PM, Giles Coochey wrote: I've tended to find that when a card is failing the MAC address starts setting itself to FF:FF:FF:FF:FF:FF not 00:00:00:XX:XX:XX FF:FF:FF:FF:FF:FF is broadcast. Sorry... in order to qualify my statement I should say that I was in charge of United Kingdom RMAs for a Taiwanese manufacturer of Network cards for 3 years... as such I have come across thousands of faulty network cards and one of our tests for a faulty PROM was to ask our distributors if the MAC addresses had set itself to FF:FF:FF:FF:FF:FF. I don't doubt that it is a broadcast address. ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On Wed, Oct 13, 2010 at 2:19 PM, Brunner, Brian T. bbrun...@gai-tronics.com wrote: I just tried a full powerdown with NTP deactivated. The system came up, the time is fine. Not that the time on the motherboard should necessarily affect the MAC on an expansion card, but that was a good test nonetheless. I'm suspicious (as others have suggested) the card itself is bad. I think to suspect the flash chip that stores the MAC addr. The rest of the card may be perfect. Using it long-term might require no more than a manual edit of the init script for it adding something to this effect 'if MAC == zeros; then set MAC to 00:0a:cd:1a:c1:71 fi'. This will fix this card without clobbering it's successor down the road. *** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This footnote also confirms that this email message has been swept for the presence of computer viruses. www.Hubbell.com - Hubbell Incorporated** Brian, While those suspicions were well justified I am not sure your guess is correct in this particular case as I just swapped the NIC I had for a different one and I seem to be getting the same sort of errors again. What's the likelihood that two NICs in a row have a faulty flash? ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
2010/10/13 Boris Epstein borepst...@gmail.com: On Wed, Oct 13, 2010 at 2:19 PM, Brunner, Brian T. bbrun...@gai-tronics.com wrote: I just tried a full powerdown with NTP deactivated. The system came up, the time is fine. Not that the time on the motherboard should necessarily affect the MAC on an expansion card, but that was a good test nonetheless. I'm suspicious (as others have suggested) the card itself is bad. I think to suspect the flash chip that stores the MAC addr. The rest of the card may be perfect. Using it long-term might require no more than a manual edit of the init script for it adding something to this effect 'if MAC == zeros; then set MAC to 00:0a:cd:1a:c1:71 fi'. This will fix this card without clobbering it's successor down the road. *** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This footnote also confirms that this email message has been swept for the presence of computer viruses. www.Hubbell.com - Hubbell Incorporated** Brian, While those suspicions were well justified I am not sure your guess is correct in this particular case as I just swapped the NIC I had for a different one and I seem to be getting the same sort of errors again. What's the likelihood that two NICs in a row have a faulty flash? Well, you can set new mac address also manually on ifcfg-ethX script .. or ifconfig .. -- Eero ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On Wed, Oct 13, 2010 at 2:55 PM, Eero Volotinen eero.voloti...@iki.fi wrote: 2010/10/13 Boris Epstein borepst...@gmail.com: On Wed, Oct 13, 2010 at 2:19 PM, Brunner, Brian T. bbrun...@gai-tronics.com wrote: I just tried a full powerdown with NTP deactivated. The system came up, the time is fine. Not that the time on the motherboard should necessarily affect the MAC on an expansion card, but that was a good test nonetheless. I'm suspicious (as others have suggested) the card itself is bad. I think to suspect the flash chip that stores the MAC addr. The rest of the card may be perfect. Using it long-term might require no more than a manual edit of the init script for it adding something to this effect 'if MAC == zeros; then set MAC to 00:0a:cd:1a:c1:71 fi'. This will fix this card without clobbering it's successor down the road. *** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This footnote also confirms that this email message has been swept for the presence of computer viruses. www.Hubbell.com - Hubbell Incorporated** Brian, While those suspicions were well justified I am not sure your guess is correct in this particular case as I just swapped the NIC I had for a different one and I seem to be getting the same sort of errors again. What's the likelihood that two NICs in a row have a faulty flash? Well, you can set new mac address also manually on ifcfg-ethX script .. or ifconfig .. -- Eero ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos OK... how do I set it? Or, more importantly, how do I find out what MAC the card currently thinks it has? Boris. ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
2010/10/13 Boris Epstein borepst...@gmail.com: On Wed, Oct 13, 2010 at 2:55 PM, Eero Volotinen eero.voloti...@iki.fi wrote: 2010/10/13 Boris Epstein borepst...@gmail.com: On Wed, Oct 13, 2010 at 2:19 PM, Brunner, Brian T. bbrun...@gai-tronics.com wrote: I just tried a full powerdown with NTP deactivated. The system came up, the time is fine. Not that the time on the motherboard should necessarily affect the MAC on an expansion card, but that was a good test nonetheless. I'm suspicious (as others have suggested) the card itself is bad. I think to suspect the flash chip that stores the MAC addr. The rest of the card may be perfect. Using it long-term might require no more than a manual edit of the init script for it adding something to this effect 'if MAC == zeros; then set MAC to 00:0a:cd:1a:c1:71 fi'. This will fix this card without clobbering it's successor down the road. *** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. This footnote also confirms that this email message has been swept for the presence of computer viruses. www.Hubbell.com - Hubbell Incorporated** Brian, While those suspicions were well justified I am not sure your guess is correct in this particular case as I just swapped the NIC I had for a different one and I seem to be getting the same sort of errors again. What's the likelihood that two NICs in a row have a faulty flash? Well, you can set new mac address also manually on ifcfg-ethX script .. or ifconfig .. -- Eero ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos OK... how do I set it? Or, more importantly, how do I find out what MAC the card currently thinks it has? Well, ifconfig? It really doesn't matter, just generate random one.. edit /etc/sysconfig/network-scripts/ifcfg-ethX Remove HWADDR=00:00:00:00:00:00 add MACADDR=WH:AT:YO:UW:AN:T0 It should work this way and then just service network restart .. -- Eero -- Eero ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
Eero Volotinen wrote: 2010/10/13 Boris Epstein borepst...@gmail.com: On Wed, Oct 13, 2010 at 2:55 PM, Eero Volotinen eero.voloti...@iki.fi wrote: 2010/10/13 Boris Epstein borepst...@gmail.com: On Wed, Oct 13, 2010 at 2:19 PM, Brunner, Brian T. bbrun...@gai-tronics.com wrote: snip I'm suspicious (as others have suggested) the card itself is bad. I think to suspect the flash chip that stores the MAC addr. The rest snip While those suspicions were well justified I am not sure your guess is correct in this particular case as I just swapped the NIC I had for a different one and I seem to be getting the same sort of errors again. What's the likelihood that two NICs in a row have a faulty flash? I'd start worrying about the m/b slot. Have you tried a different one, if one's available? Well, you can set new mac address also manually on ifcfg-ethX script .. or ifconfig .. OK... how do I set it? Or, more importantly, how do I find out what MAC the card currently thinks it has? Well, ifconfig? It really doesn't matter, just generate random one.. edit /etc/sysconfig/network-scripts/ifcfg-ethX Remove HWADDR=00:00:00:00:00:00 add MACADDR=WH:AT:YO:UW:AN:T0 It should work this way and then just service network restart .. I think I'd leave the first three octets alone - that's just the manufacturer's code. Oh, and I doubt any non-hex chars would work mark ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On Wed, Oct 13, 2010 at 12:28:27PM -0400, Boris Epstein wrote: Hi all, ... Any idea what all of this mess could mean? http://bugs.centos.org/view.php?id=4317 Tru -- Tru Huynh (mirrors, CentOS-3 i386/x86_64 Package Maintenance) http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xBEFA581B pgpSY6mz097Yp.pgp Description: PGP signature ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On 10/13/2010 1:55 PM, Eero Volotinen wrote: I just tried a full powerdown with NTP deactivated. The system came up, the time is fine. Not that the time on the motherboard should necessarily affect the MAC on an expansion card, but that was a good test nonetheless. I'm suspicious (as others have suggested) the card itself is bad. I think to suspect the flash chip that stores the MAC addr. The rest of the card may be perfect. Using it long-term might require no more than a manual edit of the init script for it adding something to this effect 'if MAC == zeros; then set MAC to 00:0a:cd:1a:c1:71 fi'. This will fix this card without clobbering it's successor down the road. While those suspicions were well justified I am not sure your guess is correct in this particular case as I just swapped the NIC I had for a different one and I seem to be getting the same sort of errors again. What's the likelihood that two NICs in a row have a faulty flash? Well, you can set new mac address also manually on ifcfg-ethX script .. or ifconfig .. But the ifcfg-ethX scripts don't run if the HWADDR entry doesn't match the NIC MAC. How do you get the right name connected to the right nic so you can even run ifconfig sensibly? -- Les Mikesell lesmikes...@gmail.com ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On 10/13/2010 1:11 PM, Les Mikesell wrote: But the ifcfg-ethX scripts don't run if the HWADDR entry doesn't match the NIC MAC. How do you get the right name connected to the right nic so you can even run ifconfig sensibly? You don't *have* to use HWADDR in the ifcfg-* file. Just comment it out on the NIC that is having problems. -- Benjamin Franz ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] the wandering MAC?
On 10/13/2010 3:16 PM, Jerry Franz wrote: On 10/13/2010 1:11 PM, Les Mikesell wrote: But the ifcfg-ethX scripts don't run if the HWADDR entry doesn't match the NIC MAC. How do you get the right name connected to the right nic so you can even run ifconfig sensibly? You don't *have* to use HWADDR in the ifcfg-* file. Just comment it out on the NIC that is having problems. Whenever I've done that on machines with multiple NICs, they've all been renamed with a .bak extension and ignored, including the one I needed to be able to reach the machine at all. Could be that there were other changes at the same time contributing to this and maybe it will work if only one doesn't match. -- Les Mikesell lesmikes...@gmail.com ___ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos