[SSI-users] Re:Linux crashes during boot up !!!

[Manimay B. <ma...@cd...>]

Hi David,
I have got the prob. solved. It was my fault in installing Linux. 
Earlier I  made a number of partitions, whereas, Linux was to be 
installed on a single partition (as the SSI guide dictates). I made it 
so and this time the kernel booted normally and mounted the shared disk 
successfully. But, now I can see a new problem cropped up - uptill now, 
the cluster has only the first node. It hangs after a few minutes of 
working. Even the console stops working, no telnet or ftp into the box 
is possible. But it responds if pinged. In this situation, I am not 
being able to add the second node to the cluster. I could not see 
anything significant in /var/log after hard rebooting it. Is this a 
known problem and we have a fix ?

The other thing I noticed was the version of OpenGFS I installed. I went 
for the one provided with SSI(v0.9.6)  distro and its version is 0.0.92 
(opengfs-0.0.92-dz). From the site of OpenGFS I found that the latest 
version of it was 0.2. My question is that, cannot we go for the latest 
(0.2) of OpenGFS with OpenSSI v0.9.6 ? May be the problem what I am 
facing now will not occur with OpenGFS 0.2 ?
Please reply
Manimay

David B. Zafman wrote:

>
> Can you send the stack trace?  I haven't played with openGFS latetly, 
> so I can't say that it has worked with the latest SSI version.
>
> Manimay Biswas wrote:
>
>> hi,
>> recently I had downloaded ssi-linux-2.4.18-v0.9.6 and fixed a kernel 
>> linux-2.4.18 with that. The  first bootup went fine except a few 
>> minor probs ( like symbolic link of chroot in /sbin etc.). Then, I 
>> configured openGFS on it by following the steps found in the 
>> installation guide. Had the / copied onto the raid. After following 
>> all the steps written in the guide, I rebooted the box with the new 
>> ramdisk. the kernel started out well in the begining but stumbled 
>> down with the following message :-
>>
>> Mounting local filesystems :  Unable to handle kernel NULL pointer 
>> dereference at virtual address 00000018
>>    *pde = 00000000
>>    Oops : 0002
>>    CPU : 1
>>     EIP : 0010 : [<c0174ad9> ]  Not tainted
>>    EFLAGS : 00010286
>>    eax : 00000000 ebx: f635a40  ecx : f6353680
>>    edx : f881e048  esi : f7cea3e0 edi : f6353680
>>    ebp : f60c3ef0   esp : f60c3edc
>>    ds : 0018 es : 0018 ss : 0018
>>    Process mount (pid : 65907, stack page = f60c30000)
>> ..........
>> ..........
>>
>>  Looking at the stack trace it seems that, the kernel panicks when it 
>> tries to execute the line -
>> action $"Mounting local filesystems: " mount -a -t nonfs,smbfs,ncpfs 
>> -O no_netdev in the /etc/rc.d/rc.sysinit file. But the best part of 
>> it is that, the kernel  does not crash every time I bring it up. But 
>> most of the time it does.
>>
>> Is it possibly a problem with the way I set it up (but, I always 
>> tried to stick to the guidelines given in the documents) ? or it has 
>> got something to do with the command 'mount' itself ?
>>
>> Please help and I can supply you with more details if required.
>> TIA ...
>> Manimay
>>
>>
>>
>>
>> -------------------------------------------------------
>> This sf.net email is sponsored by:ThinkGeek
>> Welcome to geek heaven.
>> http://thinkgeek.com/sf
>> _______________________________________________
>> Ssic-linux-users mailing list
>> Ssi...@li...
>> https://lists.sourceforge.net/lists/listinfo/ssic-linux-users
>>
>
>
>