#archiveteam 2011-11-19,Sat

↑back Search

Time Nickname Message
00:03 πŸ”— Nemo_bis hmm: $ grep -l "ERROR contacting tracker" *.log | wc -l -> 44
00:10 πŸ”— alard Nemo_bis: Same for you then, please run a git pull to help spread the load over multiple tracker instances.
00:10 πŸ”— Marcelo Why do I keep getting "no usernames avaliable" when using dld-streamer?
00:11 πŸ”— Cameron_D What is the reccomended way of running multiple download isntances?
00:12 πŸ”— db48x Cameron_D: either run multiple instances of dld-client.sh in different terminals, or run dld-streamer.sh
00:13 πŸ”— alard And, all of you: please run git pull to get more tracker names.
00:14 πŸ”— closure heh, ok I'm hitting all 3 now
00:14 πŸ”— alard Thanks. That should help, I hope. If not, I'll add some more.
00:17 πŸ”— Nemo_bis alard, so, touch STOP to complete the previous streamer and then restart?
00:17 πŸ”— alard Nemo_bis: If you can, yes. But just upgrading dld-single helps too (it takes care of half the requests to the tracker).
00:18 πŸ”— Nemo_bis ok, did it a while ago
00:19 πŸ”— alard If you get too many errors when requesting new usernames it makes sense to restart dld-stream.sh
00:25 πŸ”— Nemo_bis yes, but currently I'm running a modified version which ignores failure
00:26 πŸ”— Nemo_bis when I'm running out of disk space, what's the best way to delete what I've already uploaded to batcave?
00:27 πŸ”— db48x delete?
00:27 πŸ”— Nemo_bis from my disk
00:29 πŸ”— alard If you're feeling lucky, you could add this line to the upload-finished.sh script (belof the line with -avz --partial)
00:29 πŸ”— alard --remove-source \
00:29 πŸ”— alard --remove-source-files \
00:29 πŸ”— Nemo_bis is it dangerous? :)
00:29 πŸ”— alard (Sorry, use the last one.)
00:30 πŸ”— alard Well, if you have the space it's better to keep a copy until everything is safe on archive.org.
00:32 πŸ”— Nemo_bis but from the man I couldn't understand: in «This tells rsync to remove from the sending side the files (meaning non-directories) that are a part of the transfer and have been successfully duplicated on the receiving side», "that are a part of the transfer" means that files uploaded in previous rsyncs won't be deleted, right?
00:33 πŸ”— alard Yes, that's correct.
00:33 πŸ”— Nemo_bis :(
00:34 πŸ”— Nemo_bis well, I guess it's easier to free up some space
00:34 πŸ”— alard But if you include everything in the transfer, they will be deleted.
00:34 πŸ”— alard They may already on the receiving end, but they *are* a part of the transfer.
00:35 πŸ”— Nemo_bis oh
00:37 πŸ”— alard closure: Are you sure you've upgraded your downloaders with the new trackers? There's still a lot passing through the old address.
00:38 πŸ”— alard (Or maybe that's stuff that was already running.)
00:40 πŸ”— alard Especially your 192. ip is still active on the old one.
03:29 πŸ”— underscor chronomex: With dld-streamer, you can't dynamically change how many are running, can you?
04:12 πŸ”— chronomex no, you can't. I don't like it either, it's on the list
04:12 πŸ”— chronomex change of plans, I won't get to write any patches tonight probably
04:14 πŸ”— chronomex i really need to forward inbound ssh.
04:33 πŸ”— yipdw arg, jeez, US splinder
04:34 πŸ”— yipdw fix your damn reverse proxies
10:57 πŸ”— PepsiMax I was the last one downloading something off http://anyhub.net/
10:57 πŸ”— PepsiMax here: http://anyhub.heroku.com/
11:08 πŸ”— PepsiMax Also, the wiki needs imagemagick... It keeps bitching about creating thumbnail
12:41 πŸ”— DoubleJ yipdw: Yeah, I had to ^C at least a dozen of my dld-clients this morning since they were stuck in try-die-delete-retry loops. All USes, I think.
12:42 πŸ”— DoubleJ Maybe we could add a counter to the downloader, so after the 5th(?) try it just says screw it and moves on?
13:05 πŸ”— Nemo_bis I have 27290 MiB of incomplete downloads :-/
15:18 πŸ”— db48x DoubleJ: yea, I've got a couple of those as well
16:11 πŸ”— * db48x sighs
16:11 πŸ”— db48x hard drives are so expensive now
17:10 πŸ”— closure zsh: job table full
17:10 πŸ”— closure error messages I've never seen before :)
17:10 πŸ”— db48x lol
17:11 πŸ”— closure apparently 1000 jobs is the limit
17:20 πŸ”— Nemo_bis :-O
17:20 πŸ”— Nemo_bis and I got "No more usernames available. Entering reap mode..." on both updated dld-streamer
17:34 πŸ”— kennethre closure: that's a good problem to have
20:25 πŸ”— closure alard: there is code in dld-single.sh that seems was intended to only retry 5 times, but it's broken and retries forever on some us splinders. need to set should_retry=0 when tries -ge 5
20:31 πŸ”— closure chronomex: you wrote dld-streamer right? It seems that sometimes the tracker does not respond with a username (overloaded) and it goes into reap mode, unfortunatly
21:32 πŸ”— PepsiMax bitches love rapemode
21:44 πŸ”— Coderjoe whee
21:44 πŸ”— * Coderjoe searches through a week of logs
21:48 πŸ”— PepsiMax fine trolling: https://en.wikipedia.org/wiki/List_of_Linux_kernel_names
22:03 πŸ”— closure we've done 666666 splinder users, and are past the 50% point
22:07 πŸ”— DoubleJ OHHHHHHHHHHHHHH We're halfway theeeeeerrrrrrrre...
23:42 πŸ”— SketchCow Huzzzaaahhhh
23:49 πŸ”— SketchCow Do I need to make the call for more people, closure ?
23:51 πŸ”— closure closes on 24th.. I haven't run the numbers but we could not finish
23:54 πŸ”— db48x at 5k users per hours we'll need another 4.5 days

irclogger-viewer