I havent built a kernel on a 1&1 box yet, but Im almost positive I can do it now. Until I read that, I assumed it was purely because they use xfs, which is not available in default rh9 kernels. Thats some interesting output though, and Im wondering what the configuration of that system is. Since if Im right, and they dont work because of the xfs support then the kernel never would have gotten as far that output is indicating (8139too is a network card btw). Lack of xfs support would have caused a VFS panic.
Reading the below tells me 1) the file system mounted (maybe that system is not running xfs?) 2) the system is unavailable because the network isnt working
Either way, if someone can get me a 1&1 box I can test a kernel on, I can probably put out an rpm that would work with it.
Hmm, I have the same problem. Here is how my problem happened. I clicked on auto-update in plesk, asked me to update kernal and a few other things, I said ok. Then I couldn't log in to my webserver, so I logged into shell and decided to reboot. After the reboot, chaos.
here is my bootlog:
Dec 13 23:00:18 u15151609 syslog: syslogd startup succeeded
Dec 13 23:00:18 u15151609 syslog: klogd startup succeeded
Dec 13 23:00:18 u15151609 keytable: Loading keymap:
Dec 13 23:00:18 u15151609 keytable:
Dec 13 23:00:18 u15151609 keytable: Loading system font:
Dec 13 23:00:18 u15151609 keytable: Couldnt get a file descriptor referring to the console
Dec 13 23:00:18 u15151609 keytable:
Dec 13 23:00:18 u15151609 rc: Starting keytable: succeeded
Dec 13 23:00:19 u15151609 random: Initializing random number generator: succeeded
Dec 13 23:00:19 u15151609 netfs: Mounting other filesystems: succeeded
Dec 13 23:00:19 u15151609 autofs: automount startup succeeded
Dec 13 23:00:17 u15151609 sysctl: net.ipv4.ip_forward = 0
Dec 13 23:00:17 u15151609 sysctl: net.ipv4.conf.default.rp_filter = 1
Dec 13 23:00:17 u15151609 sysctl: kernel.sysrq = 0
Dec 13 23:00:17 u15151609 sysctl: kernel.core_uses_pid = 1
Dec 13 23:00:17 u15151609 network: Setting network parameters: succeeded
Dec 13 23:00:17 u15151609 network: Bringing up loopback interface: succeeded
Dec 13 23:00:20 u15151609 named: named startup succeeded
Dec 13 23:00:21 u15151609 sshd: succeeded
Dec 13 23:00:24 u15151609 xinetd: xinetd startup succeeded
Dec 13 23:00:24 u15151609 ntpd: failed
Dec 13 23:00:24 u15151609 ntpd: ntpd startup succeeded
Dec 13 23:00:26 u15151609 mysqld: Starting MySQL: succeeded
Dec 13 23:00:26 u15151609 courier-imap: Starting imapd succeeded
Dec 13 23:00:26 u15151609 courier-imap: Starting imap-ssl succeeded
Dec 13 23:00:27 u15151609 courier-imap: Starting pop3 succeeded
Dec 13 23:00:27 u15151609 courier-imap: Starting pop3-ssl succeeded
Dec 13 23:00:28 u15151609 qmail: Starting qmail: succeeded
Dec 13 23:00:36 u15151609 spamassassin: spamd startup succeeded
Dec 13 23:00:47 u15151609 psa-spamassassin: Starting psa-spamassassin service: succeeded
Dec 13 23:00:48 u15151609 stunnel: Starting stunnel: succeeded
Dec 13 23:00:50 u15151609 httpd: [Mon Dec 13 23:00:50 2004] [warn] module frontpage_module is already loaded, skipping
Dec 13 23:00:51 u15151609 httpd: [Mon Dec 13 23:00:51 2004] [warn] VirtualHost 217.160.249.199:80 overlaps with VirtualHost 217.160.249.199:80, the first has precedence, perhaps you need a NameVirtualHost directive
Dec 13 23:00:52 u15151609 httpd: httpd startup succeeded
Dec 13 23:00:54 u15151609 psa-spamassassin: Starting psa-spamassassin service: succeeded
Dec 13 23:00:57 u15151609 psa: Starting Plesk: failed
Dec 13 23:00:59 u15151609 crond: crond startup succeeded
Dec 13 23:01:00 u15151609 xfs: xfs startup succeeded
Dec 13 23:01:00 u15151609 anacron: anacron startup succeeded
Dec 13 23:01:00 u15151609 atd: atd startup failed
Dec 13 23:01:03 u15151609 rc: Starting mailman: succeeded
these are the lines of interest:
Dec 13 23:00:24 u15151609 ntpd: failed
Dec 13 23:00:24 u15151609 ntpd: ntpd startup succeeded
Dec 13 23:00:57 u15151609 psa: Starting Plesk: failed
Dec 13 23:01:00 u15151609 atd: atd startup failed
I can access rescue mode and mount my drives, just not sure what data to back up. There's vhosts directory and a mysqldump, what else? Any help is appreciated.
I can access rescue mode and mount my drives, just not sure what data to back up. There's vhosts directory and a mysqldump, what else? I'd like to keep the plesk settings and all mail, how can I back these up?
Id have to see that system first hand to fix it, I dont think its as bad as other problems Ive seen at 1&1 in the past. They may have stopped being so reliant on custom kernels, generally you get a kernel panic and you'd never even get anything in the boot log. (since the fs never mounts). That error you just posted looks more like something going on in userland thats breaking just a couple of apps.
I had the problem with the network card on upgrading Plesk on my 1 & 1 server.
1 & 1 support were useless, first time I'd called them, and I had to more or less talk them through using linux ....
I ended up fixing this by connecting through the serial console, and configuring the new hardware on bootup, can't remember exactly what I did as it's about 2 months since I did it, but I'm sure any non-noob admin can work it out
Heh - 1&1 and Kernels. The number of times I've managed to lock myself out of my system - you'd think I'd learn. Ahem. Anyway, I've kind of gone and done it again (oneday I'll backup /boot somewhere safe, I really will) - my whole server is now rather dodgily running off of the RAMDisk rescue thingy-majiggy. Joy.
Scott - if you still need to test a kernel on a 1&1 box drop me a line. Like I say, the thing is limping along in rescue mode right now (I've fixed this before by messing around with various kernel rpms and whatnot, but my memory is crap and I'm a little busy right now), but you'd be doing me a big favour and probably helping a few other people in the future too.
I noticed that they had set the systems up to use a serial console, but Ive never seen where you can access the terminal server. Care to fill us in on what you did to get to that?
scott wrote:I noticed that they had set the systems up to use a serial console, but Ive never seen where you can access the terminal server. Care to fill us in on what you did to get to that?
I think that Thought_Memory means that they went to the 1&1 control panel, select server Administration, select recovery tool, select "rescue system" and then "reset," which uses (I presume) a serial connection to reboot the server and tell it to boot off of a network rescue disk. Apparently on some servers (not mine) there is also a true serial console.
Chrisc: Try mounting the partition that /boot is on (/dev/hda1 [ext3] on my 1&1 system), and then do ln -s <old kernel name> vmlinuz (My 1&1 kernel is 2.4.25-040218).
You can get new 1&1 built kernels offf of the 1&1 update server, at update.onlinehome-server.com/local-updates (HTTP or anon FTP, but only from a 1and1 net [not from your home computer]).
