Hello,
new dumps (as of 20050516) available at http://download.wikimedia.org/ Please note, we no longer provide 2GB split files. I do sincerely believe, that HTTP has enough capability to do partial downloads of files ;-) And those splits simply double disk space requirements...
Cheers, Domas
Domas Mituzas wrote:
new dumps (as of 20050516) available at http://download.wikimedia.org/ Please note, we no longer provide 2GB split files. I do sincerely believe, that HTTP has enough capability to do partial downloads of files ;-) And those splits simply double disk space requirements...
Yay. I've always been wondering what sort of people would need the split files.
For Windows users, I recommend GetRight to download the file in chunks and with interruptions.
Timwi wrote:
new dumps (as of 20050516) available at http://download.wikimedia.org/ Please note, we no longer provide 2GB split files. I do sincerely believe, that HTTP has enough capability to do partial downloads of files ;-) And those splits simply double disk space requirements...
Yay. I've always been wondering what sort of people would need the split files.
We started providing split files simply because Apache on the servers they were running on would refuse to *serve* files larger than 2 gigabytes.
If your OS is unable to handle files of that size, you're probably going to have problems with the database too. :)
-- brion vibber (brion @ pobox.com)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Timwi schrieb:
For Windows users, I recommend GetRight to download the file in chunks and with interruptions.
What, you recommend mucho $$$ software when Free Download Manager [1] is available?
Honestly, I *did* buy GetRight some years ago, but scrapped it for FDM.
Magnus
[1] http://www.freedownloadmanager.org/
Magnus Manske wrote:
Timwi schrieb:
For Windows users, I recommend GetRight to download the file in chunks and with interruptions.
What, you recommend mucho $$$ software
Uhm... no?... I recommend something which is free at least in the sense of "free beer".
Haven't heard of that yet -- will try that later. I'll send my complaints about it to you via private mail, mwahaha :-)
use curl to download large files on windows.
http://curl.haxx.se/download.html
On 5/17/05, Timwi timwi@gmx.net wrote:
Magnus Manske wrote:
Timwi schrieb:
For Windows users, I recommend GetRight to download the file in chunks and with interruptions.
What, you recommend mucho $$$ software
Uhm... no?... I recommend something which is free at least in the sense of "free beer".
Haven't heard of that yet -- will try that later. I'll send my complaints about it to you via private mail, mwahaha :-)
Wikitech-l mailing list Wikitech-l@wikimedia.org http://mail.wikipedia.org/mailman/listinfo/wikitech-l
Domas Mituzas:
new dumps (as of 20050516) available at http://download.wikimedia.org/ Please note, we no longer provide 2GB split files. I do sincerely believe, that HTTP has enough capability to do partial downloads of files ;-) And those splits simply double disk space requirements...
The server doesn't handle requests for partial content correctly.
This is the output of wget -d -c http://download.wikimedia.org/wikipedia/de/old_table.sql.gz after I interrupted the previous connection at a file size of 3907378719 Bytes.
-----------------------------------------------------------------------
DEBUG output created by Wget 1.9+cvs-stable (Red Hat modified) on linux-gnu.
--09:53:24-- http://download.wikimedia.org/wikipedia/de/old_table.sql.gz => `old_table.sql.gz' Auflösen des Hostnamen »download.wikimedia.org«.... 207.142.131.212 Caching download.wikimedia.org => 207.142.131.212 Verbindungsaufbau zu download.wikimedia.org[207.142.131.212]:80... verbunden. Created socket 3. Releasing 0x94a34c0 (new refcount 1). ---request begin--- GET /wikipedia/de/old_table.sql.gz HTTP/1.0 User-Agent: Wget/1.9+cvs-stable (Red Hat modified) Host: download.wikimedia.org Accept: */* Connection: Keep-Alive Range: bytes=3907378719-
---request end--- HTTP Anforderung gesendet, warte auf Antwort... HTTP/1.0 206 Partial Content Connection: keep-alive Date: Wed, 18 May 2005 07:51:15 GMT Content-Length: 12856795601 ETag: 1881960765 Accept-Ranges: bytes Content-Type: application/x-gzip Content-Range: bytes 2147483647-15004279247/15004279248 Server: Wikimedia dump service 20050409 (lighttpd)
Found download.wikimedia.org in host_name_addresses_map (0x94a34c0) Registered fd 3 for persistent reuse. Closing fd 3 Releasing 0x94a34c0 (new refcount 1). Invalidating fd 3 from further reuse. Erneuter Versuch.
-------------------------------------------------------------------
Obviously "Range" in the request and "Content-Range" in the response don't match. This results in wget trying to request the same over and over again. (Download doesn't start at all. The last line above, "Erneuter Versuch", means "next try".) Maybe this confused the T-Online proxies as well yesterday.
My version of wget is the one that comes with Fedora Core 3 and claims to be able to handle files larger than 2GB.
Is a new stats run in progress?
Alfio
On Mon, 16 May 2005, Domas Mituzas wrote:
Hello,
new dumps (as of 20050516) available at http://download.wikimedia.org/ Please note, we no longer provide 2GB split files. I do sincerely believe, that HTTP has enough capability to do partial downloads of files ;-) And those splits simply double disk space requirements...
Cheers, Domas
Yes.
On 5/18/05, Alfio Puglisi puglisi@arcetri.astro.it wrote:
Is a new stats run in progress?
Alfio
On Mon, 16 May 2005, Domas Mituzas wrote:
Hello,
new dumps (as of 20050516) available at http://download.wikimedia.org/ Please note, we no longer provide 2GB split files. I do sincerely
believe,
that HTTP has enough capability to do partial downloads of files ;-) And those splits simply double disk space requirements...
Cheers, Domas
Wikitech-l mailing list Wikitech-l@wikimedia.org http://mail.wikipedia.org/mailman/listinfo/wikitech-l
wikitech-l@lists.wikimedia.org