| 1 |
Collin Starkweather wrote: |
| 2 |
> On Mon, 2004-03-08 at 21:06, Mark Knecht wrote: |
| 3 |
> |
| 4 |
>>>!!! ERROR: There appears to be FILE SYSTEM CORRUPTION. A file that |
| 5 |
>>>is listed |
| 6 |
>> |
| 7 |
>> This is a known bug with hashed btrees in the ext3 filesystem. It's |
| 8 |
>>been around a while. Don't know why it isn;t getting fices unless the |
| 9 |
>>fix is difficult. |
| 10 |
> |
| 11 |
|
| 12 |
I don't read gentoo-user so I didn't read the whole thread. But I |
| 13 |
suppose you are speaking about the recent perl compilation problem. If |
| 14 |
not, please ignore this email. |
| 15 |
|
| 16 |
> |
| 17 |
> This has got to be the worst error message I've seen in some time. It |
| 18 |
> is basically a panic inducer where there is no need for one. |
| 19 |
> |
| 20 |
> I have a question for those who know a thing or two about filesystems |
| 21 |
> (which I do not count myself amongst). I have a server that I won't |
| 22 |
> have physical access to for at least a few weeks. Unmounting /var and |
| 23 |
> tune2fsing as recommended in the bug report would be a major pain and I |
| 24 |
> would like to wait until I have physical access to the box just in |
| 25 |
> case. |
| 26 |
|
| 27 |
Well, someone on the forums described the following as an easier way to |
| 28 |
emerge perl without tune2fs stuff: |
| 29 |
|
| 30 |
emerge -B perl |
| 31 |
emerge -K perl |
| 32 |
|
| 33 |
It worked for me. I know it doesn't solve the whole problem, but it's a |
| 34 |
hack to get your perl working without tune2fs. |
| 35 |
|
| 36 |
> |
| 37 |
> Is this something that can cause cascading problems or can I put it off |
| 38 |
> for a few weeks? (So far the box has been functioning perfectly.) Any |
| 39 |
> thoughts / advice would be appreciated. |
| 40 |
> |
| 41 |
|
| 42 |
I don't think this problem influences the FS deeply. I've done couple of |
| 43 |
times 'shutdown -Fr now' and then tried to recompile the perl and |
| 44 |
everytime the only thing that the fsck repaired was one i-node (which |
| 45 |
had 1 reference more that it should have or something). The point is |
| 46 |
that this bug seems to only create one corrupted i-node and from what |
| 47 |
I've learned about ext3 this shouldn't affect the whole system at all. |
| 48 |
|
| 49 |
Hope it helps. |
| 50 |
|
| 51 |
--jd |
| 52 |
|
| 53 |
> -Collin |
| 54 |
> |
| 55 |
> P.S. I asked the same question of the bug responders on Bugzilla but |
| 56 |
> so far have gotten nada. |
| 57 |
> |
Archives