Grrr Debian....

Post #312,312 by folkert 8/8/09 12:31:38 PM Reply	Grrr Debian.... Ok... I have a machine 2.5 hours away from me. I had a filesystem error the other day (Thursday). So, I did a "telinit 1"... damn I forgot Debian enforces that and kills all existing connections (different from CentOS/RHEL etc...) and prompt for console password on the console. So, I call the colo hosting and they reboot the machine. (damn 517 days uptime gone) Rebooted it. Doesn't come back. I ask them to go look... seems eth0 doesn't exist anymore. FEH! Drive 2.5 hours, 3.25 with construction traffic, even though I took an "Alternate Route", it had construction traffic too, GRR. Get to the machine and do an "ifconfig -a" uno:~# ifconfig -a eth2 Link encap:Ethernet HWaddr 00:17:a4:3f:43:84 BROADCAST MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:0 (0.0 B) TX bytes:0 (0.0 B) Interrupt:19 eth3 Link encap:Ethernet HWaddr 00:17:a4:3f:43:85 BROADCAST MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:0 (0.0 B) TX bytes:0 (0.0 B) Interrupt:16 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:56943 errors:0 dropped:0 overruns:0 frame:0 TX packets:56943 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:27504232 (26.2 MiB) TX bytes:27504232 (26.2 MiB) GOSH... how in the HECK is a remote administrator to depend on this crap? That is just CRAP! Bug being filed against SOMETHING.
Post #312,320 by static 8/8/09 7:10:40 PM Reply	Maybe udev spat the dummy? Or a card's MAC changed. I saw something like that when I re-arranged NICs. However, since the last time I did that on a Linux machine, udev had come along and MAC addresses now get remembered. Wade. Q:Is it proper to eat cheeseburgers with your fingers? A:No, the fingers should be eaten separately.
Post #312,321 by Andrew Grygus 8/8/09 7:23:32 PM Reply	Yup, got bit by that udev thing over a year ago . . . I think the first time was when Network Manager decided to reconfigure stuff. Cost me a few hours (I'd never heard of udev before) but the fix instructions have been in my config manual since then.
Post #312,322 by scoenye 8/8/09 8:17:48 PM Reply	/etc/udev/rules.d/70-persistent-net.rules Take a look in there and see how the MAC addresses compare to the ifconfig results. ctime may be interesting too.
Post #312,323 by folkert 8/8/09 9:23:00 PM Reply	Re: /etc/udev/rules.d/70-persistent-net.rules Huh, here it is, this is funny, looky here: # PCI device 14e4:1659 (tg3) ACTION=="add", SUBSYSTEM=="net", DRIVERS=="?", SYSFS{address}=="00:17:a4:3f:43:85", NAME="eth0" # PCI device 14e4:1659 (tg3) ACTION=="add", SUBSYSTEM=="net", DRIVERS=="?", SYSFS{address}=="00:17:a4:3f:43:84", NAME="eth1" # PCI device 0x14e4:0x1659 (tg3) SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?", ATTR{address}=="00:17:a4:3f:43:84", ATTR{type}=="1", KERNEL=="eth", NAME="eth2" # PCI device 0x14e4:0x1659 (tg3) SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?", ATTR{address}=="00:17:a4:3f:43:85", ATTR{type}=="1", KERNEL=="eth", NAME="eth3" Notice anything? BTW, ctime was sometime on 2006-09-27 and its was rebooted sometime in February of 2008 last, mtime was about 2 months ago, until yesterday.
Post #312,325 by static 8/8/09 10:58:31 PM 8/9/09 12:03:11 AM Reply	Ouch. That looks like an upgrade to the udev files manager that perhaps included an incompatible file format. Or maybe it just forgot to read the existing file first. :-/ Well, now you can find out what package to file the bug against! :-) Wade. Q:Is it proper to eat cheeseburgers with your fingers? A:No, the fingers should be eaten separately. Edited by static Aug. 9, 2009, 12:03:11 AM EDT Expand All History
Post #312,326 by scoenye 8/9/09 12:15:36 AM Reply	SYSFS match key has disappeared The SYSFS match key has been replaced with the ATTR key. When the box rebooted, the first two lines no longer worked so the rule generator kicked in and wrote the new rules starting with the first available interface number. But I can't find exactly when SYSFS was dropped. Just various bits and pieces hinting that SYSFS had been deprecated, documentation with just one word changed, etc.
Post #312,328 by folkert 8/9/09 9:04:43 AM Reply	Oh yes, wise one. I've found the same amount of information. This should have been a MAJOR announcement in the packages notifications, during upgrades. When a major and possibly critical system is seriously changed and could possibly make things "unavailable". Oh well, I guess that is what you get for running SID on a server.

Welcome to IWETHEY!