Thread: Re: [SSI-users] Kernel Panic
Brought to you by:
brucewalker,
rogertsang
From: John B. <joh...@hp...> - 2003-07-22 19:55:42
|
-------- Original Message -------- Subject: Re: [SSI-users] Kernel Panic Date: Tue, 22 Jul 2003 12:51:13 -0700 From: John Byrne <joh...@hp...> To: Vikram <vi...@sy...> CC: ssi...@li... References: <003301c34f81$60f68510$7001a8c0@winvcg> <3F1...@hp...> <3F1...@hp...> <003d01c3500b$37e199a0$7001a8c0@winvcg> Vikram wrote: > Hi John, > > ----- Original Message ----- > From: "John Byrne" <joh...@hp...> > To: "Vikram" <vi...@sy...> > Cc: <ssi...@li...> > Sent: Tuesday, July 22, 2003 1:23 AM > Subject: Re: [SSI-users] Kernel Panic > > > >>John Byrne wrote: >> >>>Vikram wrote: >>> >>> >>>>Hi All, >>>> >>>>After creating OpenSSi initrd image i tried booting into that. During >>>>booting i got this error: >>>><error> >>>>............. >>>>............. >>>>Configuring cluster >>>>Running pre-root cluster initialization. >>>>RTNL: assertion failed at devinet.c (797) >>>>RTNL:assertion failed at devinet.c (797) >>>> >>>>Instruction(i) breakpoint #0 at 0xc0124060 (adjusted) >>>>0xc0124060 panic: int3 >>>> >>>>Entering kdb(current =0xcb8a6000, pid 40) on processor 0 due to >>>>Breakpoint >>>>@0xc0124060 >>>>[0]kdb> >>>> >>>></error> >>>> >>>>On doing 'bt' i got this: >>>><bt output> >>>>0xcb8d7f10 0xc0124060 panic (oxc03471a0, 0x0, 0x0) >>>> kernel .txt 0xc0100000 0xc0124060 0xc0 124170 >>>>0xc0221b22 ics_seticsinfo+0x92(0x0, 0xc04f7bfc, 0xcb8a6000, 0x0) >>>> kernel .txt 0xc0100000 0xc0221a90 0xc0221b40 0xcb8a7f28 >>>>oxco224e45 >>>>ics_llinit+0xef(oxco4f7740,0x0,0x7e) >>>>..............ics_init+0x5c >>>>...............cluster_main_init_preroot+0x73 >>>>...............ssisys_cluster_init_preroot+0x8 >>>> >>>></bt output> >>>> >>>>that was the output of bt. >>>> >>>>Is there is a solution for this problem ? It will be great help to me. >>>> >>>>ps: I am using Redhat 8.0 >>>> >>>>Regards >>>>Vikram >>> >>> >>>It is a little difficult to tell without the text of the panic message, >>>but it looks like that node number "0" was passed to the routine. This >>>is probably because the net drivers were not loaded correctly. When >>>executing the cluster_mkinitrd did you specify the "--with" option with >>>the NIC drivers. (If you have multiple kinds of NICs, include all the >>>drivers in the order in /etc/modules.conf.) >>> >>>John Byrne >>> >>> >> >>Bruce forwarded your other mail to me. (And I acutally bothered to read > > it.) > >>Your problem looks to be that the winbond-840.o driver depends on the >>mii.o driver being loaded first. So you need to specify "--with mii >>--with winbond-840" when you make the initrd. >> >>If this doesn't help, I'll need to know the results of an "ifconfig -a" >>and the linuxrc and etc/bootab from the initrd. (To look in an initrd: >>zcat initrd >/tmp foo; mount -o loop /tmp/foo /mnt.) >> >>John Byrne >> > > > Thanks for looking at my problem, i have included the > "--with=mii --with=winbond-840" > while creating initrd image. I am sending my "linuxrc", "boottab" and output > of "ifconfig -a" > > <linuxrc> > > #!/bin/bash > > export PATH=/bin > > echo Mounting /proc > mount -nt proc none /proc > > echo "Loading jbd module" > insmod /lib/jbd.o > echo "Loading ext3 module" > insmod /lib/ext3.o > echo "Loading mii module" > insmod /lib/mii.o > echo "Loading winbond-840 module" > insmod /lib/winbond-840.o > > echo Gathering cluster info > for iface in `ifconfig -a | grep HWaddr | > sed 's/\(eth[0-9]*\).*HWaddr \(.*\)/\1-\2/'` > do > dev=`echo $iface | cut -f 1 -d -` > addr=`echo $iface | cut -f 2 -d -` > rec=`tail +2 /etc/boottab | grep $addr` > > if [ $? -eq 0 ] > then > break > fi > done > > master=`head -1 /etc/boottab` > addr=`echo $rec | cut -f 3 -d /` > node=`echo $rec | cut -f 1 -d /` > > cat >/etc/cluster.conf <<EOF-RC > IFCONFIG=$dev:$addr > CLUSTER_MASTER=$master > CLUSTER_NODENUM=$node > EOF-RC > > echo Configuring cluster > cluster_config --prep > echo Running pre-root cluster initialization > cluster_config --preroot > > echo Mounting root in linuxrc > mount_remote_root /sysroot > if [ $? -ne "0" ] > then > # Get root device name from clustertab for specific node > # to allow multi-path disks to be mounted by whoever is up first. > cfs_mount -r ext3 /dev/hda1 /sysroot > fi > echo Unmounting /proc > doumount /proc > > echo Attempting pivot_root > cd /sysroot > pivot_root /sysroot /sysroot/initrd > echo Running post-root cluster initialization > /sbin/cluster_config --postroot > echo Starting init > exec /initrd/bin/chroot . /sbin/cluster_config --initproc </dev/console > >>/dev/console 2>&1 > > > </linuxrc> > > <boottab> > > interface=ARRAY(0x811f24c)=00:80:48:B5:27:8F > 00:80:48:B5:27:8F > 1:192.168.3.1 > 1/00:80:48:B5:27:8F/192.168.3.1:255.255.255.0//dev/hda1 > > </boottab> > > <ifconfig -a> > > eth0 Link encap:Ethernet HWaddr 00:80:48:B5:27:8F > inet addr:192.168.1.195 Bcast:192.168.1.255 Mask:255.255.255.0 > UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 > RX packets:117 errors:0 dropped:0 overruns:0 frame:0 > TX packets:48 errors:0 dropped:0 overruns:0 carrier:0 > collisions:0 txqueuelen:100 > RX bytes:10133 (9.8 Kb) TX bytes:4787 (4.6 Kb) > Interrupt:10 Base address:0x2000 > > lo Link encap:Local Loopback > inet addr:127.0.0.1 Mask:255.0.0.0 > UP LOOPBACK RUNNING MTU:16436 Metric:1 > RX packets:10 errors:0 dropped:0 overruns:0 frame:0 > TX packets:10 errors:0 dropped:0 overruns:0 carrier:0 > collisions:0 txqueuelen:0 > RX bytes:700 (700.0 b) TX bytes:700 (700.0 b) > > </ifconfig -a> > > I think the above output can be helpful for you to debug the problem, if you > need any other thing tell me, I will send it accross. > Once again thanks for looking into my problem. > > Regards > Vikram > > Unfortunately, nothing is leaping out at me here. The only thing interesting is that you changed the IP address, but I don't think that can be the problem. Look for error messages before the panic. Maybe the drivers still aren't loading. (Some other dependency I didn't see.) Is there anyway you can capture the output as the system boots? (Serial console, perhaps.) Alternatively, you can edit the linuxrc script to add debugging. (Such as catting /etc/cluster.conf or having "ifconfig -a" output.) Sorry I can't be much help, yet. John |
From: John B. <joh...@hp...> - 2003-07-22 20:09:17
|
Vikram, The "interface=" line you show in the bootab shouldn't be there (the line beginning "1:" should be the first line of the file) and is sufficient to break things, It looks like one of our perl scripts is doing something wrong, but we don't see where this is possible. I'd be interested in what version of perl you are running. The command that generates the boottab is "/sbin/mkdhcpd.conf -b". This will output what should be the boottab to standard out. I'd be interested in knowing if the interface line is output by this command. John >> <boottab> >> >> interface=ARRAY(0x811f24c)=00:80:48:B5:27:8F >> 00:80:48:B5:27:8F >> 1:192.168.3.1 >> 1/00:80:48:B5:27:8F/192.168.3.1:255.255.255.0//dev/hda1 >> >> </boottab> >> John Byrne wrote: > > > -------- Original Message -------- > Subject: Re: [SSI-users] Kernel Panic > Date: Tue, 22 Jul 2003 12:51:13 -0700 > From: John Byrne <joh...@hp...> > To: Vikram <vi...@sy...> > CC: ssi...@li... > References: <003301c34f81$60f68510$7001a8c0@winvcg> > <3F1...@hp...> <3F1...@hp...> > <003d01c3500b$37e199a0$7001a8c0@winvcg> > > Vikram wrote: > >> Hi John, >> >> ----- Original Message ----- From: "John Byrne" <joh...@hp...> >> To: "Vikram" <vi...@sy...> >> Cc: <ssi...@li...> >> Sent: Tuesday, July 22, 2003 1:23 AM >> Subject: Re: [SSI-users] Kernel Panic >> >> >> >>> John Byrne wrote: >>> >>>> Vikram wrote: >>>> >>>> >>>>> Hi All, >>>>> >>>>> After creating OpenSSi initrd image i tried booting into that. During >>>>> booting i got this error: >>>>> <error> >>>>> ............. >>>>> ............. >>>>> Configuring cluster >>>>> Running pre-root cluster initialization. >>>>> RTNL: assertion failed at devinet.c (797) >>>>> RTNL:assertion failed at devinet.c (797) >>>>> >>>>> Instruction(i) breakpoint #0 at 0xc0124060 (adjusted) >>>>> 0xc0124060 panic: int3 >>>>> >>>>> Entering kdb(current =0xcb8a6000, pid 40) on processor 0 due to >>>>> Breakpoint >>>>> @0xc0124060 >>>>> [0]kdb> >>>>> >>>>> </error> >>>>> >>>>> On doing 'bt' i got this: >>>>> <bt output> >>>>> 0xcb8d7f10 0xc0124060 panic (oxc03471a0, 0x0, 0x0) >>>>> kernel .txt 0xc0100000 0xc0124060 0xc0 124170 >>>>> 0xc0221b22 ics_seticsinfo+0x92(0x0, 0xc04f7bfc, 0xcb8a6000, 0x0) >>>>> kernel .txt 0xc0100000 0xc0221a90 0xc0221b40 0xcb8a7f28 >>>>> oxco224e45 >>>>> ics_llinit+0xef(oxco4f7740,0x0,0x7e) >>>>> ..............ics_init+0x5c >>>>> ...............cluster_main_init_preroot+0x73 >>>>> ...............ssisys_cluster_init_preroot+0x8 >>>>> >>>>> </bt output> >>>>> >>>>> that was the output of bt. >>>>> >>>>> Is there is a solution for this problem ? It will be great help to me. >>>>> >>>>> ps: I am using Redhat 8.0 >>>>> >>>>> Regards >>>>> Vikram >>>> >>>> >>>> >>>> It is a little difficult to tell without the text of the panic message, >>>> but it looks like that node number "0" was passed to the routine. This >>>> is probably because the net drivers were not loaded correctly. When >>>> executing the cluster_mkinitrd did you specify the "--with" option with >>>> the NIC drivers. (If you have multiple kinds of NICs, include all the >>>> drivers in the order in /etc/modules.conf.) >>>> >>>> John Byrne >>>> >>>> >>> >>> Bruce forwarded your other mail to me. (And I acutally bothered to read >> >> >> it.) >> >>> Your problem looks to be that the winbond-840.o driver depends on the >>> mii.o driver being loaded first. So you need to specify "--with mii >>> --with winbond-840" when you make the initrd. >>> >>> If this doesn't help, I'll need to know the results of an "ifconfig -a" >>> and the linuxrc and etc/bootab from the initrd. (To look in an initrd: >>> zcat initrd >/tmp foo; mount -o loop /tmp/foo /mnt.) >>> >>> John Byrne >>> >> >> >> Thanks for looking at my problem, i have included the >> "--with=mii --with=winbond-840" >> while creating initrd image. I am sending my "linuxrc", "boottab" and >> output >> of "ifconfig -a" >> >> <linuxrc> >> >> #!/bin/bash >> >> export PATH=/bin >> >> echo Mounting /proc >> mount -nt proc none /proc >> >> echo "Loading jbd module" >> insmod /lib/jbd.o >> echo "Loading ext3 module" >> insmod /lib/ext3.o >> echo "Loading mii module" >> insmod /lib/mii.o >> echo "Loading winbond-840 module" >> insmod /lib/winbond-840.o >> >> echo Gathering cluster info >> for iface in `ifconfig -a | grep HWaddr | >> sed 's/\(eth[0-9]*\).*HWaddr \(.*\)/\1-\2/'` >> do >> dev=`echo $iface | cut -f 1 -d -` >> addr=`echo $iface | cut -f 2 -d -` >> rec=`tail +2 /etc/boottab | grep $addr` >> >> if [ $? -eq 0 ] >> then >> break >> fi >> done >> >> master=`head -1 /etc/boottab` >> addr=`echo $rec | cut -f 3 -d /` >> node=`echo $rec | cut -f 1 -d /` >> >> cat >/etc/cluster.conf <<EOF-RC >> IFCONFIG=$dev:$addr >> CLUSTER_MASTER=$master >> CLUSTER_NODENUM=$node >> EOF-RC >> >> echo Configuring cluster >> cluster_config --prep >> echo Running pre-root cluster initialization >> cluster_config --preroot >> >> echo Mounting root in linuxrc >> mount_remote_root /sysroot >> if [ $? -ne "0" ] >> then >> # Get root device name from clustertab for specific node >> # to allow multi-path disks to be mounted by whoever is up first. >> cfs_mount -r ext3 /dev/hda1 /sysroot >> fi >> echo Unmounting /proc >> doumount /proc >> >> echo Attempting pivot_root >> cd /sysroot >> pivot_root /sysroot /sysroot/initrd >> echo Running post-root cluster initialization >> /sbin/cluster_config --postroot >> echo Starting init >> exec /initrd/bin/chroot . /sbin/cluster_config --initproc </dev/console >> >>> /dev/console 2>&1 >> >> >> >> </linuxrc> >> >> <boottab> >> >> interface=ARRAY(0x811f24c)=00:80:48:B5:27:8F >> 00:80:48:B5:27:8F >> 1:192.168.3.1 >> 1/00:80:48:B5:27:8F/192.168.3.1:255.255.255.0//dev/hda1 >> >> </boottab> >> >> <ifconfig -a> >> >> eth0 Link encap:Ethernet HWaddr 00:80:48:B5:27:8F >> inet addr:192.168.1.195 Bcast:192.168.1.255 >> Mask:255.255.255.0 >> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 >> RX packets:117 errors:0 dropped:0 overruns:0 frame:0 >> TX packets:48 errors:0 dropped:0 overruns:0 carrier:0 >> collisions:0 txqueuelen:100 >> RX bytes:10133 (9.8 Kb) TX bytes:4787 (4.6 Kb) >> Interrupt:10 Base address:0x2000 >> >> lo Link encap:Local Loopback >> inet addr:127.0.0.1 Mask:255.0.0.0 >> UP LOOPBACK RUNNING MTU:16436 Metric:1 >> RX packets:10 errors:0 dropped:0 overruns:0 frame:0 >> TX packets:10 errors:0 dropped:0 overruns:0 carrier:0 >> collisions:0 txqueuelen:0 >> RX bytes:700 (700.0 b) TX bytes:700 (700.0 b) >> >> </ifconfig -a> >> >> I think the above output can be helpful for you to debug the problem, >> if you >> need any other thing tell me, I will send it accross. >> Once again thanks for looking into my problem. >> >> Regards >> Vikram >> >> > > Unfortunately, nothing is leaping out at me here. The only thing > interesting is that you changed the IP address, but I don't think that > can be the problem. > > Look for error messages before the panic. Maybe the drivers still aren't > loading. (Some other dependency I didn't see.) Is there anyway you can > capture the output as the system boots? (Serial console, perhaps.) > > Alternatively, you can edit the linuxrc script to add debugging. (Such > as catting /etc/cluster.conf or having "ifconfig -a" output.) > > Sorry I can't be much help, yet. > > John > > > |
From: John B. <joh...@hp...> - 2003-07-22 20:12:47
|
John Byrne wrote: > > Vikram, > > The "interface=" line you show in the bootab shouldn't be there (the > line beginning "1:" should be the first line of the file) and is > sufficient to break things, It looks like one of our perl scripts is > doing something wrong, but we don't see where this is possible. I'd be > interested in what version of perl you are running. > > The command that generates the boottab is "/sbin/mkdhcpd.conf -b". This > will output what should be the boottab to standard out. I'd be > interested in knowing if the interface line is output by this command. If not, then try rebuilding the initrd and see if the line is in the boottab on the ramdisk. If it isn't, try it and see if things work. So far, this is looking very strange. John |