Corrupted OS

[Garrett Cooper]

Drew Jenkins wrote:
> I go to run /usr/sbin/sysinstall. It brings up a little GUI and asks
 > me to select. I selected post-installation configuration, and it sent
 > me back to a prompt! So I tried again, selecting the recommended
 > configuration to start over again, and it again sent me back to a
 > prompt! Besides, this is kinda dangerous. Got another, perhaps more
 > complex but *safer* way to determine if it's ufs1 or 2?
> 
> 2Also, what are softupdates and why do I need them?
> TIA,
> Drew
> 
> Garrett Cooper <youshi10 at u.washington.edu> wrote: Drew Jenkins wrote:
>> /etc/fstab says ufs. Is there a better way to check if its ufs2?
>> Drew2
>>
>> Garrett Cooper  wrote: On Mar 16, 2007, at 7:34 PM, Drew Jenkins wrote:
>>
>>> How large is "large"? Why filesystem are you using with what  
>>> options?The MySQL database was just under a gigabyte, and the Zope  
>>> Data.fs file/database was somewhere under 2 gigabytes. Options? No  
>>> options. I had symlinks from where these dbases were supposed to  
>>> live on the SCSI drives to the 500 GB drive. Then suddenly, poof!  
>>> They were gone.
>>> Drew
>> Well, I was curious because I thought it could be something to deal  
>> with the 2GB file limit. You still haven't answered my question about  
>> the filesystem though: are you using UFS2 or something else?
>>
>> Thanks,
>> -Garrett
> 
> The easiest way to figure out if you're running UFS2 is to go to the 
> disk label feature within sysinstall, and define a mount point for the 
> slice. Make sure _not_ to make any changes though as you'll be thrusting 
> yourself in the middle of a system upgrade (CTRL-C is your friend).
> 
> If it's ufs1, it should definitely be converted to ufs2. There were some 
> serious limitations in ufs1, in particular dealing with file size (2GB 
> limit I believe) and features. Someone else on the list might be able to 
> advise you or point you in the right direction if you want more details..
> 
> Also, you should be running softupdates. If not you're playing a risky 
> game of russian roulette with your data, where if corrupted things can 
> disappear between reboots if you didn't power down the machine properly 
> (power down via ATX dead man power switch, power loss, etc).
> 
> If all else fails and you're not running ufs1 on the disk, try upgrade 
> your bios or firmware controller that the disk is operating on, and get 
> back to us with more details.
> 
> Cheers,
> -Garrett

In order to get to disk label without installing from scratch, go to 
Configure -> Label. Then select your Disk, press Ok. Once the next 
window comes up, press "M" and select a mount point for the slice. Then 
look off to the right and see what version of UFS the slice is using.

Another (maybe safer?) way to do this is to run /sbin/tunefs -p 
/dev/{disk+slicename}. See if something like...

tunefs: soft updates: (-n)                                 disabled

... pops up. I used my / slice as an example, so soft updates are 
automatically disabled for it (I think this has to deal with single user 
mode and fsck?).

A short description of softupdates is available here: 
<http://en.wikipedia.org/wiki/Softupdates>, and you should read the 2nd 
reference if you want more detailed info about them.

Also, could you please bottom post. Top posting is hard to read and 
bottom-posting is the defacto standard on the FreeBSD lists.

Thanks,
-Garrett