Bug 12682 - Random 100% cpu utilization makes 7.4 difficult-impossible to use on Boom & other SB3 based devices.
: Random 100% cpu utilization makes 7.4 difficult-impossible to use on Boom & o...
Status: CLOSED WORKSFORME
Product: Logitech Media Server
Classification: Unclassified
Component: Player UI
: 7.4.0
: PC MacOS X 10.5
: P2 major with 3 votes (vote)
: 7.6.0
Assigned To: Andy Grundman
: SQLite
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2009-07-05 21:39 UTC by Caleb Crome
Modified: 2011-05-09 09:07 UTC (History)
6 users (show)

See Also:
Category: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Caleb Crome 2009-07-05 21:39:16 UTC
The latest 7.4 software on OSX periodically goes to 100% cpu utilization.  The behavior seems to be sort of random, but happens frequently.  Often when I want to scroll through a big list it happens, or even just adjusting volume.  The SB UI locks up and prevents any use.

There's nothing happening in server.log or scanner.log at that time, and perl is taking 100% CPU.

This is a critical issue for me.  Don't know if others are seeing it.  Could this be because of my large music collection?  100k tracks?

This was never a problem with 7.3.  Could this be due to a bad MAC address on one my booms?  I just noticed I have a bad one.

Squeezebox Server Status

Version: 7.4 - r27417 @ Sat Jul 4 04:01:43 PDT 2009
Hostname: macmini-5.local
IP: 192.168.1.55
HTTP Port: 9000
OS: Mac OS X 10.5.7 (9J61) - EN - utf8
Platform: x86
Perl Version: 5.8.8 - darwin-thread-multi-2level
Total Players Recognized: 6
 
Library Statistics

Total Tracks: 105,547
Total Albums: 8,606
Total Artists: 7,946
Total Genres: 291
Total Playing Time: 7689:17:47
 
Player Information

Information on all identified devices connected to Squeezebox Server
 
Henrys

Player Model: boom
Firmware: 47
Player IP Address: 192.168.1.121
Player MAC Address: 00:04:20:ff:ff:01
Wireless Signal Strength: 73%
 
Kitchen

Player Model: boom
Firmware: 47
Player IP Address: 192.168.1.120
Player MAC Address: 00:04:20:1e:xx:xx
Wireless Signal Strength: 68%
 
Sadies

Player Model: boom
Firmware: 47
Player IP Address: 192.168.1.122
Player MAC Address: 00:04:20:1e:xx:xx
Wireless Signal Strength: 88%
 
Sewing

Player Model: boom
Firmware: 47
Player IP Address: 192.168.1.127
Player MAC Address: 00:04:20:1e:xx:xx
Wireless Signal Strength: 88%  
 
Squeezebox BB1

Player Model: bb
Firmware: 7.4-r6320
Player IP Address: 192.168.1.125
Player MAC Address: 00:04:20:26:xx:xx
 
Squeezebox TT1

Player Model: tt
Firmware: 7.4-r6269
Player IP Address: 192.168.1.129
Player MAC Address: 00:04:20:22:01:7d
 
Cache Folder
/Users/crome/Library/Caches/Squeezebox Server
Preferences Folder
/Users/crome/Library/Application Support/Squeezebox Server
Plugin Folders
/Users/crome/Library/Caches/Squeezebox Server/InstalledPlugins/Plugins, /Users/crome/Library/Application Support/Squeezebox Server/Plugins, /Library/Application Support/Squeezebox Server/Plugins, /Library/PreferencePanes/Squeezebox Server.prefPane/Contents/server/Plugins
Squeezebox Server Log File

/Users/crome/Library/Logs/Squeezebox Server/server.log
 
Scanner Log File

/Users/crome/Library/Logs/Squeezebox Server/scanner.log
Comment 1 James Richardson 2009-07-07 12:23:03 UTC
Caleb: does this happen when the scanner starts?
Comment 2 Andy Grundman 2009-07-07 12:41:55 UTC
Enable scan.scanner and scan.auto.
Comment 3 A.C. Marek 2009-07-22 06:02:07 UTC
Thought I'd jump in here, though not sure how helpful my comments will be.  I thought I'd audition 7.4 last night (OS X 10.5.7, Boom, ~11k tracks).  Scanning is generally quite quick on my MacBook Pro, but I figured I'd let it alone and let it scan while I slept.  

Woke up this morning, and the scanner had frozen.  Like OP, my CPU utilization was 100%.  CPU temperature, alarmingly, was 185 degrees Fahrenheit (with my machine, normal is anywhere from 120 to 160).

This did not happen immediately when the scanner started, since I monitored it for a few minutes.  

If you need any more information from me, let me know and I'll be happy to provide it.  Thanks.
Comment 4 Caleb Crome 2009-07-22 06:31:08 UTC
This does not seem to be the scanner.  Ienabld the log options you asked for,and there is no correlation between freezes and log entries.

Anything else to try?

Bumping severity as this makes my classic based devices perform unacceptably.
Comment 5 Andreas Sandberg 2009-07-27 04:17:30 UTC
I'm experiencing very similar issues with Squeezecenter 7.3.4~27825 on a 32 bit Debian machine. Symptoms are ~100% CPU utilization and nothing out of the ordinary in the logs. Haven't found a good way of reproducing the issue, except for using the devices for some time (happens at least once a day).

From looking at the strace output it seems like the main loop is running wild. The majority of the system calls logged are gettimeofday and select calls. The select calls all have a timeout of 0.0.
Comment 6 Andreas Sandberg 2009-07-27 07:41:42 UTC
One thing I forgot to mention in my previous comment: I didn't notice this bug until maybe a week ago. Nothing has changed in my setup except for one thing, I've started to use the line in on my Squeezebox Boom. Don't know if this is related at all or just a coincidence.
Comment 7 Andy Grundman 2009-07-29 14:40:47 UTC
Moving all SQLite-related bugs to 8.0.
Comment 8 Cushing Whitney 2009-11-01 11:28:49 UTC
I'm seeing the same problem since I installed 7.4 on a x86 Fedora Core 5 server with dual 2.4 GHz Athlons. The squeezeboxserver process will get pin one CPU at 100% for minutes at a time when some interaction is attempted on the Boom. My two Duet controllers continue to work fine and don't seem to pin CPU utilization. For example, pulling up the artist list from 'My Music' take 1-2 seconds on the Duets, but upwards of a minute or two on all the Booms in the house. This is not a SQLlite issue as I'm running MySQL as the backend. Please re-target this bug to the 7.4.x series as it does compromise the usage of the Booms.


Version: 7.4.1 - r28948 @ Wed Oct 21 04:01:50 PDT 2009
Hostname: server.bitlathe.com
IP: 192.168.0.xxx
HTTP Port: 9000
OS: Red Hat - EN - utf8
Platform: i686-linux
Perl Version: 5.8.8 - i386-linux-thread-multi
MySQL Version: 5.0.27
Total Players Recognized: 6
 
Library Statistics
Total Tracks: 24,924
Total Albums: 2,247
Total Artists: 1,510
Total Genres: 110
Total Playing Time: 1759:57:34
 
Player Information
 
Bedroom
Player Model: boom
Firmware: 50
Player IP Address: 192.168.0.131
Player MAC Address: 00:04:20:1e:60:34
Wireless Signal Strength: 88%
 
Kitchen
Player Model: boom
Firmware: 50
Player IP Address: 192.168.0.132
Player MAC Address: 00:04:20:1e:5f:24
Wireless Signal Strength: 90%
 
Living Room
Player Model: receiver
Firmware: 65
Player IP Address: 192.168.0.125
Player MAC Address: 00:04:20:16:8b:db
 
Office
Player Model: receiver
Firmware: 65
Player IP Address: 192.168.0.139
Player MAC Address: 00:04:20:16:ac:ae
 
Playroom
Player Model: boom
Firmware: 50
Player IP Address: 192.168.0.100
Player MAC Address: 00:04:20:1e:0b:c3
 
Squeezebox
Player Model: boom
Firmware: 50
Player IP Address: 192.168.0.126
Player MAC Address: 00:04:20:1e:79:d4
Wireless Signal Strength: 96%
Comment 9 Fazle Khan 2009-11-28 11:22:54 UTC
I am also encountering this same problem on my WindowsXP server.  Like others it does not happen immediately, but after the server has been up for a day or more.  The server will pin the cpu and cause all the players to be unresponsive.  The only thing in the logs are timeouts from the server trying to connect to mysqueezebox.com and failing I assume because the operating system can't get clock cycles to respond to the DNS request.
Comment 10 Marc Auslander 2009-11-28 15:18:08 UTC
Are very long playlists involved?  If so, see:

bug_10889
Comment 11 Cushing Whitney 2009-11-28 15:26:36 UTC
This issue doesn't involve playlists at all for me. The CPU utilization and unresponsiveness occur when I'm just trying to access the various options under 'my music' on the boom. I have no static playlists defined at all.
Comment 12 Andy Grundman 2011-01-12 12:08:32 UTC
SQLite bugs need to go back to 7.6 target.
Comment 13 Andy Grundman 2011-01-12 12:10:11 UTC
Haven't heard of this issue in a long time, please reopen if necessary.