1 |
Nikos Chantziaras wrote: |
2 |
> Saphirus Sage wrote: |
3 |
>> This has been a consistent problem on one of my computers, to no avail |
4 |
>> thusfar, even after changing sync server settings or redigesting |
5 |
>> individual failed ebuilds. In running emerge -uavDN world, it will take |
6 |
>> at least an hour to generate a list of packages to be merged, but not |
7 |
>> before ouputting a TON of failed digests. Example below. |
8 |
>> |
9 |
>> gentoog4 proc # emerge -uavDN world |
10 |
>> |
11 |
>> These are the packages that would be merged, in order: |
12 |
>> |
13 |
>> [...] |
14 |
>> |
15 |
>> So, is this just the crying screams of a failing drive, or is there some |
16 |
>> other problem at hand here? |
17 |
> |
18 |
> Does dmesg say anything about disk I/O errors? Also, try something |
19 |
> like this: |
20 |
> |
21 |
> cp -a /usr/portage /tmp/ |
22 |
> diff -r /usr/portage /tmp/portage |
23 |
> |
24 |
> If the diff claims something differs, then it's pretty certain your |
25 |
> disk or RAM might be failing. For RAM, try memtest86+. |
26 |
> |
27 |
> |
28 |
Followed the instructions and checked with diff, absolutely no |
29 |
difference, which leads me to believe it's failing memory; but there's a |
30 |
problem. This is a PowerPC machine (eMac G4), and memtest86+ seems to be |
31 |
dependent upon an x86 architecture. So, is there any good way to test |
32 |
this, or do I just pull out one of the 512 DIMMS and hope for the best? |