2.6.8.1 kernel kills my linode
It boots just fine and dandy into 2.4 kernel.
But try to put the 2.6.8.1 kernel on there? wooo watch out.
The box wont even come up. I click on boot, the overview says its running, but if I try to use lish and connect to it, lish tells me my linode isnt running.
You may want to double check your kernel caker.
8 Replies
-Chris
@EviL_SmUrF:
So in order to run the newest kernel, I'll have to do that on the 2.4 kernel, and then have it reboot into the 2.6.8.1 kernel?
Yes. If you do a logview in lish after booting 2.6.8.1, it most likely ends with a kernel panic. More and more stuff in RedHat and related distros relies on Thread local Storage, which doesn't work (yet) under UML.
> ip_conntrack version 2.1 (463 buckets, 3704 max) - 296 bytes per conntrack
ip_tables: (C) 2000-2002 Netfilter core team
iptrecent v0.3.1: Stephen Frost <
sfrost@snowman.net >.http://snowman.net/projects/ipt recent/arp_tables: (C) 2002 David S. Miller
Initializing IPsec netlink socket
NET: Registered protocol family 1
NET: Registered protocol family 10
IPv6 over IPv4 tunneling driver
ip6_tables: (C) 2000-2002 Netfilter core team
NET: Registered protocol family 17
NET: Registered protocol family 15
Bridge firewalling registered
Ebtables v2.0 registered
802.1Q VLAN Support v1.8 Ben Greear <
greearb@candelatech.com >All bugs added by David S. Miller <
davem@redhat.com >Initializing software serial port version 1
Using anticipatory io scheduler
/dev/ubd/disc0: unknown partition table
/dev/ubd/disc1: unknown partition table
Initializing stdio console driver
md: Autodetecting RAID arrays.
md: autorun …
md: … autorun DONE.
kjournald starting. Commit interval 5 seconds
EXT3-fs: mounted filesystem with ordered data mode.
VFS: Mounted root (ext3 filesystem) readonly.
INIT: version 2.85 booting
Kernel panic: handle_trap - failed to wait at end of syscall, errno = 4, status = 4479
<6>SysRq : Show Regs
EIP: 0073:[<400007a2>] CPU: 0 Not tainted ESP: 007b:bffff934 EFLAGS: 00000206
Not tainted
EAX: 0000008c EBX: 01200011 ECX: 00000000 EDX: 00000000
ESI: 00000000 EDI: 40146528 EBP: bffff9a8 DS: 007b ES: 007b
Call Trace: [<801e7e0c>] [<8001ec46>] [<80040426>] [<80033d62>] [<80033dcb>]
[<8002264a>] [<80022918>] [<800230f0>] [<802e60d8>] [<802e6331>]
This does not happen with kernel 2.6.7, which boots normally.
This doesn't seem to be NPTL-related.
@thoth39:
This does not happen with kernel 2.6.7, which boots normally.
This doesn't seem to be NPTL-related.
Just me, jumping to conclusions again
@thoth39:
This does not happen with kernel 2.6.7, which boots normally. This doesn't seem to be NPTL-related.
I'm able to boot FC2 under 2.6.8.1 just fine:
[root@li7-3 root]# uname -a
Linux li7-3 2.6.8.1-linode4-1um #1 Sat Sep 18 22:53:39 EDT 2004 i686 i686 i386 GNU/Linux
[root@li7-3 root]# cat /etc/redhat-release
Fedora Core release 2 (Tettnang)
@thoth39:
Kernel panic: handle_trap - failed to wait at end of syscall, errno = 4, status = 4479
I've seen that bug before. The UML guys are working they way through them. It has to do with re-trying the waitpid call without just exiting if it fails the first time around. I added the incremental "eintr" patch, so this kernel is far better off than any of the other 2.6 (or even 2.4 based kernels) in this regard. You were just lucky, I guess.
Mind giving it another try?
-Chris
@EviL_SmUrF:
im gonna go ahead and leave my linode runnin on the latest 2.4 until all of these bugs are sorted through and caker says he recommends for us to use the new kernel.
Good idea. The 2.4 series is nice and stable:-)
% uname -r ; uptime
2.4.26-linode29-1um
10:24:42 up 109 days, 19:00, 1 user, load average: 0.05, 0.05, 0.01