Bugzilla – Bug 15452
7.5.0 embedded doesn't run on ubuntu server 8.04
Last modified: 2011-01-14 05:00:03 UTC
Don't get it to run on my ubuntu machine. Runs but is incredibly slow while using almost NO CPU. Web interface sometimes loads (soetimes not) but never works (e.g. doesn't get any players). I don't see a scanner but the SQLite files change so the scanner is probably running, eating up whatever (can't be CPU, see above; memory is also not an issue, as far as I can see) Scan not over after about 12h, "normal" SBS takes about 45min. I disabled all plugins, no change. logs a lot of connection errors to SN and test.SN (if there are still some plugin errors in here I turned them all off and didn't see a change).
Created attachment 6430 [details] server log file SBS embedded r29755
Still no luck @r29883. Better CPU use, Scanner completes something (get log entries) but no contact through web ui. Slim::Utils::Scanner::Local::rescan (134) Discovering files in /mnt/slugpublic/Media [10-01-24 03:18:18.1043] Slim::Plugin::MusicMagic::Common::__ANON__ (75) Failed fetching filters from MusicIP [10-01-24 03:18:38.2603] Slim::Formats::XML::gotErrorViaHTTP (332) Error: getting http://www.test.mysqueezebox.com/api/sounds/v1/opml Connect timed out: Illegal seek [10-01-24 03:18:38.2617] Slim::Plugin::Sounds::Plugin::_gotSoundsError (83) Error: Unable to cache Sounds & Effects menu from SN: Connect timed out: Illegal seek [10-01-24 03:18:38.2737] Slim::Networking::Discovery::Players::_players_error (145) Unable to get players: Connect timed out: [10-01-24 03:18:38.2761] Slim::Networking::Discovery::Players::_players_error (145) Unable to get players: Connect timed out: [10-01-24 03:18:38.2786] Slim::Networking::Discovery::Players::_players_error (145) Unable to get players: Connect timed out: [10-01-24 03:18:38.2813] Slim::Networking::Discovery::Players::_players_error (145) Unable to get players: Connect timed out: [10-01-24 03:18:48.2969] Slim::Networking::SqueezeNetwork::_error (550) Unable to login to SN: Couldn't resolve IP address for: www.test.mysqueezebox.com [10-01-24 03:18:48.2987] Slim::Networking::SqueezeNetwork::_init_error (213) Unable to login to mysqueezebox.com, sync is disabled: Couldn't resolve IP address for: www.test.mysqueezebox.com (http://www.test.mysqueezebox.com) [10-01-24 03:18:48.3009] Slim::Networking::SqueezeNetwork::_init_error (227) mysqueezebox.com sync init failed: Couldn't resolve IP address for: www.test.mysqueezebox.com, will retry in 300 (http://www.test.mysqueezebox.com) [10-01-24 03:19:50.1901] Slim::Networking::Discovery::Players::_players_error (145) Unable to get players: Connect timed out: Inappropriate ioctl for device A ping to www.test.squeezenetwork.com works fine.
Andy is there some logging information that would be useful to find this problem?
Hmm, and if you run 7.4.2 or 7.5 trunk you have no problems like this? I can setup a VM with that version of Ubuntu and try it out.
7.5.0 works like a charm.
Sh... my last comment got lost, so here it is again: I, too, have it installed on a VM with Ubuntu Server 8.04 LTS on a powerful server and it work fine there so other factors have to add to it. However, the Ubuntu installation that fails is also pretty new and clean, here are the main differences: 1. MusicIP. I heard MIP doesn't work on embedded, the non-working system uses it. Could this cause it? 2. The non-working system is a rather slow ViaC7 bases system (1GHz) so I wouldn't completely rule out scheduling/priorities and a race condition.
OK, maybe a bit more "meat". The problems seem to be related to the scanner. It's a bit faster now and I let it run a day and after 16h the scan was completed and now it runs just fine. 16h as compared to 40min with 7.5 trunk! During that time I get all times of timeouts on _anything_ that tries to use sockets: Network comms, MySB errors, my SBs disconnect every 2min or so, WebUI, you name it. Is there some limit on the number of sockets that can be used on a system? Or it's simply some race condition. As I said: it _looks_ better now since it's obviously unusable in a real-world scenario. Maybe this helps a bit...
Here's my ulimit -a output, btw: core file size (blocks, -c) 0 data seg size (kbytes, -d) unlimited scheduling priority (-e) 0 file size (blocks, -f) unlimited pending signals (-i) 7671 max locked memory (kbytes, -l) 32 max memory size (kbytes, -m) unlimited open files (-n) 1024 pipe size (512 bytes, -p) 8 POSIX message queues (bytes, -q) 819200 real-time priority (-r) 0 stack size (kbytes, -s) 8192 cpu time (seconds, -t) unlimited max user processes (-u) 7671 virtual memory (kbytes, -v) unlimited file locks (-x) unlimited
On a normal system a full wipe scan is still run as a separate process so it's not doing any network stuff, etc. Are you sure it's properly running scanner.pl? (On Touch only, scanning is done within the main process). You could also try manually running scanner.pl, either with or without the main server running.
I'm not sure it's doing anything properly at all. I did a clean install and that's what it does :) The reports I'm seeing are from the server (it's in server.log) and I'm seeing this only happens as long as the scanner is running. As soon as the scanner is finished everything works fine and the timeouts go away. I don't get any scanner logs (scanner doesn't seem to log at all) but something seems to hold up the scanner, too, otherwise I can't explain the 16h of scanning.
One more thing before I remove embedded again (due to https://bugs-archive.lyrion.org/show_bug.cgi?id=15923 still rendering it unusable for me): Could it be that embedded on ubuntu needs a lot of memory for some reason? My system is somewhat memory limited in that it has 1GB of RAM but only 256MB of swap space. I can see that my system uses ~512MB for caching and actual memory use by SBS is somewhere in the 200 MB range.
Oh, and one more thing: Did you mean separate process or thread? There was no separate scanner process.
This works in the current 7.6 betas.