Manuel Bouyer wrote:

> On Sun, Oct 30, 2005 at 12:33:38PM +0100, Stéphane Witzmann wrote:
>> Hi
>> 
>> After creating then almost filling a 160GB ext2 partition under NetBSD, I
>> ran fsck on it. It says "filesystem contains large files but lacks
>> LARGE_FILE flag in superblock; fix y/n ?". As far as I know, the
>> partition is not readable on NetBSD if this flag is set. So, it is ok to
>> say no ? Since I used NetBSD to copy the data, it would probably have
>> told me if something was wrong. I mean I don't have large files for which
>> ext2 doesn't work unless the flag is turned on... or so I hope.
>> 
>> So here is question 1 : is the data (it is my backup, I'm trying ext2 so
>> that I can also read it on Linux or Windows [there is a driver, which I
>> am gonna try soon]) safe or not ?
>> 
>> Now question 2 : every time I fscked this ext2 partition (even before I
>> put any data on it), the system hung up soon after fsck returned. The
>> first time, it hung up when I synced, and right now (after copying the
>> data) I could sync but mount didn't return.
>> 
>> $ fsck.ext2 -f /dev/wd1e
>> ...
>> $ sync
>> $ mount
>> /dev/wd0a on / type ffs (soft dependencies, local)
>> /dev/wdoe on /home type ffs (soft dependencies, local)
>> kernfs on /kern type kernfs (local)
>> 
>> At this point, I can type stuff but not switch to another tty.
>> This is not a permanent hangup, the system seems to go back to normal a
>> few minutes later.
> 
> Use the char device (/dev/rwd1e) instead. You have filled your memory with
> disk blocks, possibly swapping out active processes.
> 

Tested again (with /dev/wd1e), that doesn't look like it. Before it freezes,
top says something like :

0.5% user, 0.0% nice, 1.5% system, 85.0% interrupt, 12.0% idle

Memory: 24M Act, 244K Wired, 1424K Exec, 3944K File, 165M Free
Swap: 2048M Total, 2048M Free

The hard drive doesn't sound as if something was swapped out.