1 |
Daniel Pielmeier wrote: |
2 |
> Paul Sobey schrieb: |
3 |
>> For the last couple of weeks I've noticed that emerge -puv world takes |
4 |
>> an absolute age on my machine (currently at 5 mins and counting), |
5 |
>> while the cpu gows to 100%. The little emerge rotator bar that spins |
6 |
>> while it calculates dependencies moves extremely slowly. Is there any |
7 |
>> known bug that causes this? How can I go about troubleshooting? |
8 |
>> |
9 |
>> Cheers, |
10 |
>> Paul |
11 |
> |
12 |
> |
13 |
> My guess is you have removed the portage-cache before using emerge. |
14 |
> Normally this happens when a previous emerge --sync fails because of an |
15 |
> error or there are identical timestamps on the server and the local |
16 |
> machine. |
17 |
> |
18 |
> When this happens you have no portage-cache because portage deletes the |
19 |
> cache first when syncing. If the timestamps are identical the cache is |
20 |
> not regenerated. If there is an error while syncing the cache is missing |
21 |
> too, because the regeneration of the cache is done in the end of the sync. |
22 |
> |
23 |
> So the cache is regenerated when doing the first emerge and this takes |
24 |
> quite a while. Even on a Core2Duo 6600@2.40GHz i takes more then three |
25 |
> minutes :-). |
26 |
> |
27 |
> In case of the timestamp issue i think it is a bug [1]. I think portage |
28 |
> should first check the timestamps and if they are different it should |
29 |
> delete the cache! |
30 |
|
31 |
Interesting idea - I've been running eix-sync in my crontab. Perhaps |
32 |
that hasn't been updating the edb cache. I've changed it eix-sync -r, |
33 |
perhaps that will fix things. |
34 |
|
35 |
Paul |
36 |
|
37 |
-- |
38 |
gentoo-user@l.g.o mailing list |