RE: Speed of current server
"Dennis Pinckard" <[email protected]>
| Newsgroups | gmane.network.up2date.current.devel |
|---|---|
| Message-ID | <[email protected]> |
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Thanks for the update. I've read through the docs and knew the project was in a 'get it working' phase and not perf tuning, I just didn't expect it to be that slow. I'll adjust the ssl option. Right now, I only have the laptop to play with as a client and it keeps wanting to go into suspend mode every 20 minutes or so. I've disabled all power managment (apmd and in the BIOS), but it still wants to suspend. This is all off-topic, but might have something to do with it. I can scp and ftp to/from the laptop at normal network speeds, though. - -----Original Message----- From: Hunter Matthews [mailto:[email protected]] Sent: Wednesday, March 20, 2002 2:46 PM To: Current Server Mailing List Subject: Re: [Current-server] Speed of current server Dennis, I'm using this as an opportunity to remind the list of where we are and where we hope to be going. I'm not picking on you. In general, right now performance is pretty bad. Why? Its a single threaded pure python daemon. End of story. What can be done to fix it? Well, right now we suck the whole package into ram, and then spew out to the client. I've been meaning to take a look at that and see if chunking out to 32k or 64k would improve things. (it would defnately improve our memory usage) Second thing is go into your up2date config file in /usr/sysconfig/rhn and set the ssl option for packages to no. This helps A LOT, because now you aren't encrypting/decrypting 1-100MB of data on each throw. (If you have a highly hostile network, you don't want to do this) Third is to wait for or contribute to the planned 1.3 release, which will migrate us from this pure python daemon to apache/mod_python. On Wed, 2002-03-20 at 15:04, Dennis Pinckard wrote: > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > I finally stepped off the cliff and installed 1.0.2. This was my > first attempt and it was extremely easy. Thanks for a great tool! > > I have a switched 100Mb network internally with only 20 users or > so, yet up2date is reporting between 20 and 10k/sec on updates. Is > this normal? Just wondering if this was a python issue. > > I tested it on my laptop with a fresh RedHat 7.2 install and the > RPM directories mounted via NFS from my RH7.2 file server. I don't > remember for sure, but it seemed to perform much faster when > updating the laptop. > > Setting it up again, this time on the RedHat 7.2 file server (which > has the latest updates from RedHat applied), I'm attempting to > update the laptop and I'm only getting throughput of about 14/k > sec. > > TIA. > > Dennis Pinckard > Network Administrator > Aegis Learning Group > > -----BEGIN PGP SIGNATURE----- > Version: PGP 7.0.4 > > iQA+AwUBPJjrPXBWGat9hZ87EQJPbwCXbu4SxF47kuv1VaB/lAJ3i78ZQwCgqoAJ > 0dZOa8/Z5tdxZRm3mvT6mLA= > =9SKo > -----END PGP SIGNATURE----- > _______________________________________________ > Current-server mailing list > [email protected] > http://lists.dulug.duke.edu/mailman/listinfo/current-server > > - -- Hunter Matthews Unix / Network Administrator Office: BioScience 145/244 Duke Univ. Biology Department Key: F0F88438 / FFB5 34C0 B350 99A4 BB02 9779 A5DB 8B09 F0F8 8438 Never take candy from strangers. Especially on the internet. _______________________________________________ Current-server mailing list [email protected] http://lists.dulug.duke.edu/mailman/listinfo/current-server -----BEGIN PGP SIGNATURE----- Version: PGP 7.0.4 iQA/AwUBPJj4Q3BWGat9hZ87EQKHFQCgjgIAu2bFhxEBxXxDpUQJ6Se12kgAniFT gITdN7e7tb+SHtlUP+FdfTKy =MFE6 -----END PGP SIGNATURE-----