1 |
On 01-03-2018 12:13:14 +0100, Michael Weiser wrote: |
2 |
> Hi Fabian, |
3 |
> |
4 |
> On Thu, Mar 01, 2018 at 08:46:26AM +0100, Fabian Groffen wrote: |
5 |
> |
6 |
> > I ran a --sync on my macbook this morning: |
7 |
> |
8 |
> I synced at 11:56 from rsync2. |
9 |
|
10 |
By the way. Syncing at 11:56 means you need to finish within 25 seconds |
11 |
or so. On rsync1 this is 110 seconds. (This is the difference of a |
12 |
Core i5 4x3.4GHz versus a G5 2x2.5GHz.) |
13 |
|
14 |
So, for your purposes of getting a consistent sync, you better sync |
15 |
around 50 and 20 (assuming you can get the sync to finish within 6 |
16 |
minutes). |
17 |
|
18 |
> I get: |
19 |
> |
20 |
> # time /usr/local/gentoo/usr/portage/scripts/rsync-generation/hashgen hashverify /usr/local/gentoo/usr/portage |
21 |
> verifying /usr/local/gentoo/usr/portage... |
22 |
> RSA key fingerprint 0204 A8AB D003 E57A 9558 850D BA08 091E C631 7B3C |
23 |
> good signature made 2018-03-01 10:26:41 UTC by |
24 |
> Gentoo Prefix Portage Snapshot Signing Key (Automated Signing Key) |
25 |
> Manifest.files.gz:scripts/Manifest.gz: |
26 |
> - file size mismatch |
27 |
> got: 1948 |
28 |
> expected: 1949 |
29 |
|
30 |
Yay, rsync1 and rsync2 disagreed on the size for both files. After |
31 |
unpacking and sorting the contents, it was equal though. Lesson learnt: |
32 |
the contents should be sorted to ensure equality. |
33 |
|
34 |
Now I'll put this up so I can see if I can track the reference error (in |
35 |
Manifest.files.gz). |
36 |
|
37 |
> manifest verification failed |
38 |
> 6.36 real 4.20 user 4.12 sys |
39 |
> |
40 |
> So it looks a lot better but still not fully fixed. |
41 |
> |
42 |
> > hashverify is too fast to be correct (on a |
43 |
> > 6-core machine it does the tree in less than a second) at the moment, |
44 |
> > but I'm adding counters and shizzle to verify that it's actually doing |
45 |
> > something. |
46 |
> |
47 |
> The speed is unbelievable in the sense of the word. A quick dtruss shows |
48 |
> that it is in fact opening and reading files at least. |
49 |
|
50 |
Yes, also, I noticed that q (from portage-utils) updates the cache |
51 |
usually in below-second times, which walks the entire tree as well. So |
52 |
perhaps it is all not that impossible. When generating manifests, |
53 |
hashgen is also quite fast. Of course it's operating on a warm tree, |
54 |
but it's doing writes all the time, so... |
55 |
|
56 |
Thanks, |
57 |
Fabian |
58 |
|
59 |
-- |
60 |
Fabian Groffen |
61 |
Gentoo on a different level |