Bug 501195
Summary: | extreme memory usage in pulseaudio | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Peter F. Patel-Schneider <pfpschneider> |
Component: | pulseaudio | Assignee: | Lennart Poettering <lpoetter> |
Status: | CLOSED NOTABUG | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | medium | Docs Contact: | |
Priority: | low | ||
Version: | rawhide | CC: | lkundrak, lpoetter, wtogami |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2009-05-17 23:26:05 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
Peter F. Patel-Schneider
2009-05-17 20:23:33 UTC
What are you referring to? VSZ? That is the virtual memory size. It denotes how much address space PA has allocated, not how much memory PA actaully uses. Each PA client and PA itself share a 64mb shm segment. That might amount to a lot of address space, however only a tiny bit of those segs is usually actually ever used. ALSO, PA links against to a lot of libraries. Those all all mapped into the adress space, although only a small subset of them is actually used at a time. RSS is a an equally useless value. There are other tools (like 'exmap') that are more useful to determine memory usage of Linux processes. The RSS, which I believe is the resident memory size is also over 1GB. Are you saying that that doesn't matter at all, as well? I am aware that you can't count on these numbers being completely accurate, but surely pulseaudio shouldn't have that big a memory footprint. Right now, I'm seeing pfps 2214 0.0 0.2 101988 5208 ? S<sl 02:13 0:04 /usr/bin/pulseaud which is much better. So something was causing these numbers to grow, and I was experiencing a noticable change in responsiveness when the numbers were big, and there is the large number of messages in the log. Something was rotten in the state of pulseaudio. If exmap is a better tool, then perhaps there should be a good way to find out about it. It doesn't appear to be installed on my machine, and neither can yum find it. Looking back at /var/log/messages, here is the start of the complaints from pulseaudio May 16 16:15:45 idefix NetworkManager: <info> (wlan0): supplicant connection state: group handshake -> completed May 16 16:17:17 idefix kernel: tcsh[3591]: segfault at 0 ip 0808b90a sp bfac0320 error 4 in tcsh[8048000+54000] May 16 16:19:00 idefix pulseaudio[2124]: alsa-source.c: Increasing wakeup watermark to 15.99 ms May 16 16:22:11 idefix pulseaudio[2124]: module-rescue-streams.c: Failed to move source output 1 "PulseAudio Volume Control" to combined.monitor. May 16 16:22:13 idefix pulseaudio[2124]: alsa-source.c: Increasing minimal latency to 1.00 ms May 16 16:22:15 idefix pulseaudio[2124]: module-rescue-streams.c: Failed to move source output 6 "PulseAudio Volume Control" to combined.monitor. May 16 16:22:17 idefix pulseaudio[2124]: module-rescue-streams.c: Failed to move source output 7 "PulseAudio Volume Control" to combined.monitor. May 16 16:22:17 idefix pulseaudio[2124]: reserve-wrap.c: Failed to acquire reservation lock on device 'Audio0': Input/output error May 16 16:22:28 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:28 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:29 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:29 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:29 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:29 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:29 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:29 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:29 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:29 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:29 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:33 idefix pulseaudio[2124]: ratelimit.c: 87 events suppressed May 16 16:22:33 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:33 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:33 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:33 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:34 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:34 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:34 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:34 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:34 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:34 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:34 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:38 idefix pulseaudio[2124]: ratelimit.c: 89 events suppressed May 16 16:22:38 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:38 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:38 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:38 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:39 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally May 16 16:22:39 idefix pulseaudio[2124]: asyncq.c: q overrun, queuing locally |