View previous topic :: View next topic |
Author |
Message |
PrakashP Veteran
Joined: 27 Oct 2003 Posts: 1249 Location: C.C.A.A., Germania
|
Posted: Sun Dec 05, 2004 11:48 am Post subject: |
|
|
SQLite - see portage-c... (needs less than 1 (sometimes to about 2) seconds to find something for me, cdb needs about 4-5 seconds) But it could also be python slowing down things a lot... |
|
Back to top |
|
|
SuperJudge Apprentice
Joined: 08 Jun 2004 Posts: 177 Location: Albany, GA
|
Posted: Mon Dec 06, 2004 3:49 pm Post subject: |
|
|
emerge metadata is taking a long time. _________________ ---Powered by Honda |
|
Back to top |
|
|
s4kk3 Apprentice
Joined: 15 Oct 2004 Posts: 232 Location: Finland
|
Posted: Mon Dec 06, 2004 3:53 pm Post subject: |
|
|
yes. emerge metadata takes long but everything else is 100x faster for me. _________________ My own filemanager project |
|
Back to top |
|
|
steveb Advocate
Joined: 18 Sep 2002 Posts: 4564
|
Posted: Mon Dec 06, 2004 4:13 pm Post subject: |
|
|
s4kk3 wrote: | yes. emerge metadata takes long but everything else is 100x faster for me. | for me everything else is about 100% faster (double the speed as without cdb).
cheers
SteveB |
|
Back to top |
|
|
SuperJudge Apprentice
Joined: 08 Jun 2004 Posts: 177 Location: Albany, GA
|
Posted: Mon Dec 06, 2004 4:44 pm Post subject: |
|
|
s4kk3 wrote: | yes. emerge metadata takes long but everything else is 100x faster for me. | I'm quite certain that I've been emerging metadata for over an hour now. _________________ ---Powered by Honda |
|
Back to top |
|
|
tobidope n00b
Joined: 17 Aug 2003 Posts: 24 Location: Germany
|
Posted: Mon Dec 06, 2004 5:33 pm Post subject: |
|
|
Quote: | I'm quite certain that I've been emerging metadata for over an hour now. Shocked |
I'm sorry, but that is a portage problem like i described before. Maybe you all should try anydbm instead of cdb. If this much faster than cdb I would suggest to use that db backend. Or I should try to send a patch to the devs for enhancing the metadata code-branch |
|
Back to top |
|
|
SuperJudge Apprentice
Joined: 08 Jun 2004 Posts: 177 Location: Albany, GA
|
Posted: Mon Dec 06, 2004 5:49 pm Post subject: |
|
|
It's no problem, I just didn't realize it was supposed to take a long time.
I'll background it. _________________ ---Powered by Honda |
|
Back to top |
|
|
tobidope n00b
Joined: 17 Aug 2003 Posts: 24 Location: Germany
|
Posted: Mon Dec 06, 2004 5:58 pm Post subject: |
|
|
I didn't want to be offensive in anyway, but I will also try anydbm. And maybe I'll have to climb down in portage.py for enhancing the cache generation. I see no need for syncing after each key insertion (that's what makes emerge metatdata so slow), but maybe I should ask somebody who knows about that. Somebody here with expert portage knowledge? I think I should go to the "Portage & Pogramming" forum. |
|
Back to top |
|
|
TrueDFX Retired Dev
Joined: 02 Jun 2004 Posts: 1348
|
Posted: Mon Dec 06, 2004 9:23 pm Post subject: |
|
|
Wow, I'm impressed. I read that emerge --metadata would be slow, but this...
emerge --metadata in one terminal. emerge -1 glibc in another. emerge -1 glibc is moving the files to / right now, while emerge --metadata is at 82%.
(I'm sure it would have been a *lot* faster if I waited with the glibc installation until --metadata finished.)
As long as searching with emerge is fast afterwards, I'm happy. And it is |
|
Back to top |
|
|
mirko_3 l33t
Joined: 02 Nov 2003 Posts: 605 Location: Birreria
|
Posted: Tue Dec 07, 2004 1:17 pm Post subject: |
|
|
If searching is all it improves, I don't see why you don't just use (like I do) esearch: it's in portage, I rebuild the database nightly, and search is not fast - it's instant... _________________ Non fa male! Non fa male! |
|
Back to top |
|
|
Hydraulix Guru
Joined: 12 Dec 2003 Posts: 447
|
Posted: Wed Dec 08, 2004 10:08 am Post subject: |
|
|
Has this been updated in the new version of Portage? _________________ It is the fate of operating systems to become free.
- Neal Stephenson |
|
Back to top |
|
|
TazG Guru
Joined: 22 May 2004 Posts: 320 Location: Canada
|
Posted: Wed Dec 08, 2004 1:13 pm Post subject: |
|
|
You know I don't think it's as fast as you are saying it is. You see if you emerge -S mozilla twice, it will be faster the second time whether you are using cdb or not because it is cached in memory.
emerge -S mozilla without cdb: Code: | real 3m9.985s
user 1m25.189s
sys 0m13.098s | emerge -S mozilla AGAIN without cdb: Code: | real 1m47.062s
user 1m22.551s
sys 0m11.705s | emerge -S mozilla with cdb: Code: | real 1m19.555s
user 1m9.137s
sys 0m4.165s | So. It may look like it's 2 minutes faster if you don't do it twice without cdb, but it's actually only about 30 seconds faster. |
|
Back to top |
|
|
PrakashP Veteran
Joined: 27 Oct 2003 Posts: 1249 Location: C.C.A.A., Germania
|
Posted: Wed Dec 08, 2004 2:03 pm Post subject: |
|
|
But the difference is, cdb is fast from the first time, and even faster the second time when data files are cached... |
|
Back to top |
|
|
TazG Guru
Joined: 22 May 2004 Posts: 320 Location: Canada
|
Posted: Wed Dec 08, 2004 2:46 pm Post subject: |
|
|
Actually a fourth emerge -S mozilla (second time with cdb) is only 4 seconds faster than the third. |
|
Back to top |
|
|
tobidope n00b
Joined: 17 Aug 2003 Posts: 24 Location: Germany
|
Posted: Wed Dec 08, 2004 2:52 pm Post subject: |
|
|
And the second difference, if you have an older machine with not so much RAM and disk cache most oft the time when you do something portage related the process is waiting for I/O. Of course throwing more RAM and processorpower at something solves almost any performance problem. That's one point of view and mostly it's cheaper to buy more hardware instead of solving a problem efficient. But in my opinion the db backend of portage is something where you could think 5 minutes to enhance it. In an ideal world there should be no need for something like esearch, because the portage cache would be fast enough. |
|
Back to top |
|
|
killercow Tux's lil' helper
Joined: 29 Jan 2004 Posts: 86 Location: Netherlands
|
Posted: Thu Dec 09, 2004 1:58 pm Post subject: local mysql server as lan-wide portage dir? |
|
|
isn;t it theoretically possible to connect to any mysql server?
If this mysql server sits in the lan on my personal house server all of my other machines should be able to connect to it and search the portage tree on that machine?
How would we do this? im no python whiz (alltough i can do some stuff) so is there a quick way to do this?
Would it also suffice to just update the portage tree on the server (or does portage keep a different tree based on the emerged packages per system?) |
|
Back to top |
|
|
tobidope n00b
Joined: 17 Aug 2003 Posts: 24 Location: Germany
|
Posted: Thu Dec 09, 2004 4:47 pm Post subject: |
|
|
@killercow
Read the first post of this thread. There is a reference to the mysql solution. |
|
Back to top |
|
|
killercow Tux's lil' helper
Joined: 29 Jan 2004 Posts: 86 Location: Netherlands
|
Posted: Wed Dec 15, 2004 10:03 am Post subject: noticed the mysql part |
|
|
could it be possible to keep one portage tree in mysql on a seperate machine? And still use all different clients (my machine, my laptops,my servers, my brothers pc, etc) to use this one portage tree and still have a different set of installed programs? |
|
Back to top |
|
|
Mindphaser Tux's lil' helper
Joined: 15 Dec 2004 Posts: 113 Location: Germany
|
Posted: Wed Dec 15, 2004 10:12 pm Post subject: |
|
|
Very nice, but "emerge sync" is hell slow now....
Is it normal, that before syncing the Portage Cache, he simply deletes it ?
So "Updating Portage Cache" takes as long like "emerge metadata" does....
And, sadly, i dont get a massive speed-up like some other people here.
"emerge kde --search > bla" took 4,5sec. , with cdb-powered-portage, 4,3sec
However, thats better than nothing, but that long "Update Portage Cache" after every sync are realy disturbing....
btw: I have a Athlon XP 2600+ with 512MB RAM |
|
Back to top |
|
|
LucaSpiller Apprentice
Joined: 10 Sep 2004 Posts: 188 Location: Censorship Land (aka England)
|
Posted: Wed Dec 15, 2004 10:56 pm Post subject: |
|
|
It seems ok, there doesn't really seem to be much of a speed increase for me - mainly because I don't search that much. Searching is quite a bit quicker but is still rather slow.
Portage-c looks good so I might give that a try. _________________ :: Luca :: Mac Fag :: Original Macbook, 2g RAM :: Closet Linux user (seasoned with salt and pepper) :: C2D E4400 @ 2ghz, 4g RAM (only 3.2g detected under 64bit...), Nvidia 9600GSO :: |
|
Back to top |
|
|
gentoo_lan l33t
Joined: 08 Sep 2004 Posts: 891 Location: Charles Town, WV
|
Posted: Thu Dec 16, 2004 12:04 am Post subject: |
|
|
It looks pretty good to me so far. Good work. |
|
Back to top |
|
|
thoffmeyer Apprentice
Joined: 11 Apr 2004 Posts: 208 Location: GMT -5 Hours
|
Posted: Thu Dec 16, 2004 3:35 pm Post subject: |
|
|
works great over here, im too lazy to post but thanks. |
|
Back to top |
|
|
tobidope n00b
Joined: 17 Aug 2003 Posts: 24 Location: Germany
|
Posted: Sun Dec 19, 2004 11:00 am Post subject: |
|
|
Mindphaser wrote: | Very nice, but "emerge sync" is hell slow now....
Is it normal, that before syncing the Portage Cache, he simply deletes it ?
So "Updating Portage Cache" takes as long like "emerge metadata" does....
And, sadly, i dont get a massive speed-up like some other people here.
"emerge kde --search > bla" took 4,5sec. , with cdb-powered-portage, 4,3sec
However, thats better than nothing, but that long "Update Portage Cache" after every sync are realy disturbing....
btw: I have a Athlon XP 2600+ with 512MB RAM |
Try the new module. emerge sync should be much faster! |
|
Back to top |
|
|
PrakashP Veteran
Joined: 27 Oct 2003 Posts: 1249 Location: C.C.A.A., Germania
|
Posted: Sun Dec 19, 2004 11:17 am Post subject: |
|
|
[Edit] Forget it, I somehow missed the first line...
Yup, it's a *lot* faster on updating database. Very nice. |
|
Back to top |
|
|
matroskin Apprentice
Joined: 21 Jan 2003 Posts: 214
|
Posted: Sun Dec 19, 2004 12:10 pm Post subject: |
|
|
after upgrading to new /usr/lib/portage/pym/portage_db_cdb.py
I get this error.
Code: | root@anatolya:~ # emerge -pv system
These are the packages that I would merge, in order:
Calculating system dependencies /auxdb exception: [/usr/portage::dev-lang/python-2.1.3-r1]: 'NoneType' object has no attribute 'get'
Traceback (most recent call last):
File "/usr/bin/emerge", line 2935, in ?
if not mydepgraph.xcreate(myaction):
File "/usr/bin/emerge", line 1320, in xcreate
myeb=portage.portdb.xmatch("bestmatch-visible",mydep)
File "/usr/lib/portage/pym/portage.py", line 5363, in xmatch
myval=best(self.xmatch("match-visible",None,mydep=mydep,mykey=mykey))
File "/usr/lib/portage/pym/portage.py", line 5374, in xmatch
myval=match_from_list(mydep,self.xmatch("list-visible",None,mydep=mydep,mykey=mykey))
File "/usr/lib/portage/pym/portage.py", line 5360, in xmatch
myval=self.gvisible(self.visible(self.cp_list(mykey)))
File "/usr/lib/portage/pym/portage.py", line 5458, in gvisible
myaux=db["/"]["porttree"].dbapi.a
ux_get(mycpv, ["KEYWORDS"])
File "/usr/lib/portage/pym/portage.py", line 5117, in aux_get
if self.auxdb[mylocation][cat].has_key(pkg):
File "/usr/lib/portage/pym/portage_db_cdb.py", line 106, in has_key
if self.data.cdbObject.get(key) is not None:
AttributeError: 'NoneType' object has no attribute 'get'
|
Can u post plz initial version of this python file? |
|
Back to top |
|
|
|