[CentOS] Re: Boot Log Problems - more info

[Robert Slade]

On Sun, 2007-10-28 at 16:57 +0000, Robert Slade wrote: 
> On Sun, 2007-10-28 at 12:16 -0400, Tom Diehl wrote:
> > On Sun, 28 Oct 2007, Robert Slade wrote:
> > 
> > > On Sun, 2007-10-28 at 16:33 +0100, Alain Spineux wrote:
> > >> On 10/28/07, Robert Slade <centos at likley.co.uk> wrote:
> > >>> On Sun, 2007-10-28 at 13:02 +0100, Alain Spineux wrote:
> > >>>> I have this in my syslog.conf :
> > >>>>
> > >>>> # Save boot messages also to boot.log
> > >>>> local7.*
> > >>>
> > >>> Mine says:
> > >>>
> > >>> # Save boot messages also to boot.log
> > >>> local7.*                                                /var/log/boot.log
> > >>
> > >> Yes my cut&paste was a little lazy !
> > >>
> > >>>
> > >>>>
> > >>>> But did you check dmesg for kernel messages ?
> > >>>>
> > >>>> regards
> > >>>
> > >>>
> > >>> dmesg only lists messages from prior to update too :-(.
> > >>
> > >> dmesg display the kernel internal buffer. It should be emptied when rebooting !
> > >> It is impossible to see messages prior the last reboot !
> > >>
> > >> What about your other log file ?
> > >> /var/log/messages is filled normaly ?
> > >>
> > >>>
> > >>> Thanks for the reply
> > >>>
> > >>> Rob
> > >>>
> > >>>>
> > >>>> On 10/28/07, Robert Slade <centos at likley.co.uk> wrote:
> > >>>>> Hi,
> > >>>>>
> > >>>>> I have just updated from CentOS 4 to 5 and I am seeing a problem with
> > >>>>> udev during booting, but I am unable to track it down as it looks like
> > >>>>> the boot.log is not working. The file is there but empty and the
> > >>>>> previous log only lists events up to the point at which I upgraded.
> > >>>>>
> > >>>>> syslog.conf looks ok to me in so much as there is an entry pointing
> > >>>>> to /var/log/boot.log.
> > >>>>>
> > >>>>> Any suggestions as to what I should be looking at?
> > >>>>>
> > >>>>> Rob
> > >
> > > Both dmesg and boot.log do not have any entries from before the update.
> > >
> > > /var/log/messages appears to have normal entries ie from today but there
> > > are no error messages.
> > >
> > > I have checked syslog it is running
> > 
> > Just a thought, rpm -V sysklogd. See what you get. Also have you run ckrootkit
> > or similar on the machine?
> > 
> > I know you said this started after an update and most likely you are correct
> > but it is just a thought. You never know what kind of coincidence you might
> > run into.
> 
> rpm -V sysklogd gave nothing, but rpm -q sysklogd gave:
> 
> sysklogd-1.4.1-39.2
> 
> Ran chkrootkit - nothing came up.
> 
> Rob

After a lot of fiddling, I have managed to catch the original error I
was trying to investigate - I guess that this maybe the cause or
contributing to the missing logs.

Right after the interactive logon message (which doesn't work) the clock
is set then I get:

starting_udev: udev[485]:add_to_rules: invalid KERNEL operation
udev[485]:invalid_rule: /etc/udev/rules.d/90-ib rule:1
udev[485]:add_to_rules: invalid KERNEL operation
etc up to rule:7

The contents of 90-ib is:

KERNEL="umad*", NAME="infiniband/%k"
KERNEL="issm*", NAME="infiniband/%k"
KERNEL="ucm*", NAME="infiniband/%k", MODE="0666"
KERNEL="uverbs*", NAME="infiniband/%k", MODE="0666"
KERNEL="uat", NAME="infiniband/%k", MODE="0666"
KERNEL="ucma", NAME="infiniband/%k", MODE="0666"
KERNEL="rdma_cm", NAME="infiniband/%k", MODE="0666"

Rob