Ok I found a leak in Ape.config().
A quick fix would be to replace
APE_Server/scripts/commands/inlinepush.js by this :
http://pastebin.com/swb77hwa
Let me know.
Thanks.
Anthony C.
Le 16/06/2011 12:45, Anthony Catel a écrit :
Hi,
In order to get a valid Valgrind report you need to build spidermonkey
with "--enable-valgrind --enable-debug".
Also, the memory report can't be valid du to the garbage collection.
Anyway, there may be a leak when trying to send a message on a
non-existing channel. I'll take a look on that today.
Le 16/06/2011 12:41, Szymon a écrit :
If it helps I'm "valgrind" aped, these two things are thrown from time
to time.
==13813== Use of uninitialised value of size 8
==13813== at 0x650BEA9: js_IsAboutToBeFinalized(void*) (in /home/
project/mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x650DA14:
js::ConservativeGCStackMarker::markWord(unsigned long) (in /home/
project/mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x650F2B3: js_TraceRuntime(JSTracer*) (in /home/
project/mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x650F566: GCUntilDone(JSContext*, JSGCInvocationKind)
(in /home/project/mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x651118A: LastDitchGC(JSContext*) (in /home/project/
mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x6511761: js_NewFinalizableGCThing(JSContext*,
unsigned int) (in /home/project/mesjasz/modules/lib/
libmod_spidermonkey.so)
==13813== by 0x6587B31: js_NewString (in /home/project/mesjasz/
modules/lib/libmod_spidermonkey.so)
==13813== by 0x64C579E: JS_NewStringCopyN (in /home/project/mesjasz/
modules/lib/libmod_spidermonkey.so)
==13813== by 0x64C2AA4: ape_sm_cmd_wrapper (in /home/project/
mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x409C75: process_cmd (in /usr/local/bin/aped)
==13813== by 0x409FF1: checkcmd (in /usr/local/bin/aped)
==13813== by 0x4083A7: checkrecv (in /usr/local/bin/aped)
==13813==
==13813== Conditional jump or move depends on uninitialised value(s)
==13813== at 0x650DA17:
js::ConservativeGCStackMarker::markWord(unsigned long) (in /home/
project/mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x650F2B3: js_TraceRuntime(JSTracer*) (in /home/
project/mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x650F566: GCUntilDone(JSContext*, JSGCInvocationKind)
(in /home/project/mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x651118A: LastDitchGC(JSContext*) (in /home/project/
mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x6511761: js_NewFinalizableGCThing(JSContext*,
unsigned int) (in /home/project/mesjasz/modules/lib/
libmod_spidermonkey.so)
==13813== by 0x6587B31: js_NewString (in /home/project/mesjasz/
modules/lib/libmod_spidermonkey.so)
==13813== by 0x64C579E: JS_NewStringCopyN (in /home/project/mesjasz/
modules/lib/libmod_spidermonkey.so)
==13813== by 0x64C2AA4: ape_sm_cmd_wrapper (in /home/project/
mesjasz/modules/lib/libmod_spidermonkey.so)
==13813== by 0x409C75: process_cmd (in /usr/local/bin/aped)
==13813== by 0x409FF1: checkcmd (in /usr/local/bin/aped)
==13813== by 0x4083A7: checkrecv (in /usr/local/bin/aped)
==13813== by 0x417F5C: parser_ready_http (in /usr/local/bin/aped)
And report:
==13859== HEAP SUMMARY:
==13859== in use at exit: 518,698 bytes in 9,321 blocks
==13859== total heap usage: 260,933 allocs, 251,612 frees,
92,508,480 bytes allocated
==13859==
==13859== LEAK SUMMARY:
==13859== definitely lost: 129,210 bytes in 3,473 blocks
==13859== indirectly lost: 362,079 bytes in 5,804 blocks
==13859== possibly lost: 14,352 bytes in 13 blocks
==13859== still reachable: 13,057 bytes in 31 blocks
==13859== suppressed: 0 bytes in 0 blocks
==13859== Rerun with --leak-check=full to see details of leaked memory
==13859==
==13859== For counts of detected and suppressed errors, rerun with: -v
==13859== Use --track-origins=yes to see where uninitialised values
come from
==13859== ERROR SUMMARY: 3511 errors from 40 contexts (suppressed: 6
from 6)
On 16 Cze, 11:46, Anthony Catel<a.ca...@weelya.com> wrote:
Ok. Let me take a look
Le 16/06/2011 11:23, Szymon a crit :
Hello,
No. I'm getting many "UNKNOWN_CHANNEL".
On 16 Cze, 11:18, Anthony Catel<a.ca...@weelya.com> wrote:
Hey,
It's possible. You are gettin gmany "NOT_A_CHANNEL" ?
Le 16/06/2011 10:05, Szymon a crit :
Hi,
After 12 h of running aped ate 300 MB of memory... I think it's not
normal. :P Maybe it's related to fact, that I'm sending many
messages
to non-existent channels?
On 15 Cze, 17:14, Szymon<szy...@mwg.pl> wrote:
Hi,
I see some new messages, but not mine, so maybe it is lost. I'll
try
again, then.
I'm using APE latest git version (1.2-dev). There are only mootols
(it's mandatory AFAIK? :P) and inlinepush modules. We are using
it now
only on our test version website. There are few (15-20, 2-3
connected
to APE now) people that have access to it. Every 1 minute, there
are
10 crontab jobs, running parallel. Every job sends ~50-100
messages to
different private channels (99% of them are not reaching
destination
at all, because there are few users with access to test version
(connected to APE), as I said). Problem is, that every time job
sent
package of messages, memory of aped grows. After 1 hour of running
aped use 50 MB of avaiable memory.
It's normal?
--
You received this message because you are subscribed to the Google
Groups "APE Project" group.
To post to this group, send email to ape-project@googlegroups.com
To unsubscribe from this group, send email to
ape-project+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/ape-project?hl=en
---
APE Project (Ajax Push Engine)
Official website : http://www.ape-project.org/
Git Hub : http://github.com/APE-Project/
