Hi,
I just used checking tool to scan ganglia code and found 2 defects. Do you
think this can explain the 3G virtual memory issue? Below is detail of these 2
defects.
gmetad/gmetad.c, line 149
146 if (do_sum)
147 {
148 tt = in_type_list(type, strlen(type));
149 if (!tt) return 0;
If we returned on line 149, then storage pointed by "hash_datum" will leak.
apr-1.3.3/memory/unix/apr_pools.c, line 946
941 if ((node = allocator_alloc(pool_allocator,
942MIN_ALLOC - APR_MEMNODE_T_SIZE)) == NULL) {
943 if (abort_fn)
944abort_fn(APR_ENOMEM);
945
946 return APR_ENOMEM;
If we returned on line 946, then storage pointed by "pool_allocator" will leak.
Thanks,
Jerry
在2009-05-13 09:42:52，Jerry <ada...@126.com> 写道：
I did NOT try 3.1.2. For it may need several months to reproduce, it's hard to
provide a quick answer here. Below is the result of pmap. FYI
23867: /usr/local/ganglia/sbin/gmetad
--conf=/usr/local/ganglia/etc/ganglia/gmetad.conf
0000000000400000 52K r-x-- /usr/local/ganglia/sbin/gmetad
000000000050c000 8K rw--- /usr/local/ganglia/sbin/gmetad
000000000050e000 1043356K rw--- [ anon ]
0000000040000000 4K ----- [ anon ]
0000000040001000 8192K rw--- [ anon ]
0000000040801000 4K ----- [ anon ]
0000000040802000 8192K rw--- [ anon ]
0000000041002000 4K ----- [ anon ]
0000000041003000 8192K rw--- [ anon ]
0000000041803000 4K ----- [ anon ]
0000000041804000 8192K rw--- [ anon ]
0000000042004000 4K ----- [ anon ]
0000000042005000 8192K rw--- [ anon ]
0000000042805000 4K ----- [ anon ]
0000000042806000 8192K rw--- [ anon ]
0000000043006000 4K ----- [ anon ]
0000000043007000 8192K rw--- [ anon ]
0000000043807000 4K ----- [ anon ]
0000000043808000 8192K rw--- [ anon ]
0000000044008000 4K ----- [ anon ]
0000000044009000 8192K rw--- [ anon ]
00002aaaaaaab000 808K rw--- [ anon ]
00002aaaaac00000 132K rw--- [ anon ]
00002aaaaac21000 892K ----- [ anon ]
00002aaaaad00000 1719296K rw--- [ anon ]
00002ac4cf99f000 80K r-x-- /lib64/ld-2.3.6.so
00002ac4cf9b3000 4K rw--- [ anon ]
00002ac4cfab2000 12K rw--- /lib64/ld-2.3.6.so
00002ac4cfab5000 80K r-x-- /usr/local/ganglia/lib/libganglia-3.1.1.so.0.0.0
00002ac4cfac9000 1020K ----- /usr/local/ganglia/lib/libganglia-3.1.1.so.0.0.0
00002ac4cfbc8000 20K rw--- /usr/local/ganglia/lib/libganglia-3.1.1.so.0.0.0
00002ac4cfbcd000 4K rw--- [ anon ]
00002ac4cfbdd000 4K rw--- [ anon ]
00002ac4cfbde000 260K r-x-- /usr/lib64/librrd.so.2.0.8
00002ac4cfc1f000 1024K ----- /usr/lib64/librrd.so.2.0.8
00002ac4cfd1f000 36K rw--- /usr/lib64/librrd.so.2.0.8
00002ac4cfd28000 520K r-x-- /lib64/tls/libm-2.3.6.so
00002ac4cfdaa000 1020K ----- /lib64/tls/libm-2.3.6.so
00002ac4cfea9000 16K rw--- /lib64/tls/libm-2.3.6.so
00002ac4cfead000 72K r-x-- /lib64/libnsl-2.3.6.so
00002ac4cfebf000 1020K ----- /lib64/libnsl-2.3.6.so
00002ac4cffbe000 12K rw--- /lib64/libnsl-2.3.6.so
00002ac4cffc1000 12K rw--- [ anon ]
00002ac4cffc4000 124K r-x-- /usr/local/ganglia/lib/libexpat.so.1.5.2
00002ac4cffe3000 1020K ----- /usr/local/ganglia/lib/libexpat.so.1.5.2
00002ac4d00e2000 20K rw--- /usr/local/ganglia/lib/libexpat.so.1.5.2
00002ac4d00e7000 40K r-x-- /usr/lib64/libconfuse.so.0.0.0
00002ac4d00f1000 1020K ----- /usr/lib64/libconfuse.so.0.0.0
00002ac4d01f0000 8K rw--- /usr/lib64/libconfuse.so.0.0.0
00002ac4d01f2000 148K r-x-- /usr/local/ganglia/lib/libapr-1.so.0.3.3
00002ac4d0217000 1024K ----- /usr/local/ganglia/lib/libapr-1.so.0.3.3
00002ac4d0317000 20K rw--- /usr/local/ganglia/lib/libapr-1.so.0.3.3
00002ac4d031c000 4K rw--- [ anon ]
00002ac4d031d000 8K r-x-- /lib64/libuuid.so.1.2
00002ac4d031f000 1024K ----- /lib64/libuuid.so.1.2
00002ac4d041f000 4K rw--- /lib64/libuuid.so.1.2
00002ac4d0420000 36K r-x-- /lib64/tls/librt-2.3.6.so
00002ac4d0429000 1024K ----- /lib64/tls/librt-2.3.6.so
00002ac4d0529000 8K rw--- /lib64/tls/librt-2.3.6.so
00002ac4d052b000 60K rw--- [ anon ]
00002ac4d053a000 20K r-x-- /lib64/libcrypt-2.3.6.so
00002ac4d053f000 1020K ----- /lib64/libcrypt-2.3.6.so
00002ac4d063e000 4K rw--- /lib64/libcrypt-2.3.6.so
00002ac4d063f000 188K rw--- [ anon ]
00002ac4d066e000 8K r-x-- /lib64/libdl-2.3.6.so
00002ac4d0670000 1024K ----- /lib64/libdl-2.3.6.so
00002ac4d0770000 4K rw--- /lib64/libdl-2.3.6.so
00002ac4d0771000 60K r-x-- /lib64/tls/libpthread-2.3.6.so
00002ac4d0780000 1020K ----- /lib64/tls/libpthread-2.3.6.so
00002ac4d087f000 8K rw--- /lib64/tls/libpthread-2.3.6.so
00002ac4d0881000 16K rw--- [ anon ]
00002ac4d0885000 1064K r-x-- /lib64/tls/libc-2.3.6.so
00002ac4d098f000 1020K ----- /lib64/tls/libc-2.3.6.so
00002ac4d0a8e000 96K rw--- /lib64/tls/libc-2.3.6.so
00002ac4d0aa6000 24K rw--- [ anon ]
00002ac4d0aac000 404K r-x-- /usr/lib64/libfreetype.so.6.3.8
00002ac4d0b11000 1024K ----- /usr/lib64/libfreetype.so.6.3.8
00002ac4d0c11000 48K rw--- /usr/lib64/libfreetype.so.6.3.8
00002ac4d0c1d000 144K r-x-- /usr/lib64/libpng.so.3.26.0
00002ac4d0c41000 1020K ----- /usr/lib64/libpng.so.3.26.0
00002ac4d0d40000 16K rw--- /usr/lib64/libpng.so.3.26.0
00002ac4d0d44000 80K r-x-- /usr/lib64/libz.so.1.2.3
00002ac4d0d58000 1024K ----- /usr/lib64/libz.so.1.2.3
00002ac4d0e58000 4K rw--- /usr/lib64/libz.so.1.2.3
00002ac4d0e59000 4K rw--- [ anon ]
00002ac4d0e5a000 84K r-x-- /usr/lib64/libart_lgpl_2.so.2.3.17
00002ac4d0e6f000 1020K ----- /usr/lib64/libart_lgpl_2.so.2.3.17
00002ac4d0f6e000 12K rw--- /usr/lib64/libart_lgpl_2.so.2.3.17
00002ac4d0f71000 8K rw--- [ anon ]
00002ac4d0f83000 36K r-x-- /lib64/libnss_files-2.3.6.so
00002ac4d0f8c000 1020K ----- /lib64/libnss_files-2.3.6.so
00002ac4d108b000 8K rw--- /lib64/libnss_files-2.3.6.so
00007fffdb0f6000 84K rw--- [ stack ]
ffffffffff600000 4K r-x-- [ anon ]
total 2860736K
在2009-05-13?00:58:13，"Jesse?Becker"?<haw...@gmail.com>?写道：
>This?sounds?like?a?slow?memory?leak.??Do?you?see?it?in?version?3.1.2?as?well?
>
>2009/5/12?Jerry?<ada...@126.com>:
>>?Hi?Simon,
>>
>>?Thank?you?for?your?questions.?Below?is?my?response.
>>
>>>What?is?your?ganglia?version?
>>?????3.1.1
>>
>>>What?is?the?cluster?scale??(How?many?monitored?nodes?do?you?have?)
>>?????22?nodes
>>
>>>How?long?has?the?cluster?been?up?
>>?????The?cluster?has?been?up?for?at?least?3?monthes
>>
>>>And?what?is?the?OS?of?your?head?node?
>>?????We?use?rpath?linux?and?do?some?change?in?it.?With?"uname?-a",?you?can
>>?get?"Linux?RMG1-013?2.6.24.7-9.smp.gcc3.4.x86_64?#1?SMP?Tue?Jul?15?12:23:51
>>?EDT?2008?x86_64?x86_64?x86_64?GNU/Linux".?If?you?want?to?know?the?concrete
>>?library?version,?I?can?provide?later.
>>
>>>Does?it?always?stay?like?this?
>>?????We?don't?see?this?very?frequently.
>>
>>?Thanks,
>>
>>?Jerry
>>
>>
>>?在2009-05-12?14:38:16，"Simon?Yan"?<simonya...@gmail.com>?写道：
>>>2009/5/12?Jerry?<ada...@126.com>:
>>>>?Hi,
>>>>
>>>>?In?my?environment,?I?found?that?gmetad?consumes?nearly?3G?virtual?memory.
>>>>?Below?is?the?detail.
>>>>
>>>>?[r...@abc?~]#?top?-b?-n?1?|?grep?gmetad
>>>>?23867?nobody????20???0?2793m?1424??644?S????0??0.0???0:49.27?gmetad
>>>>
>>>>?Does?anyone?see?this?before??Any?idea?about?this?
>>>>
>>>>?Thanks,
>>>>
>>>>?Jerry
>>>>
>>>>
>>>>
>>>>?________________________________
>>>>?穿越地震带?纪念汶川地震一周年
>>>>?------------------------------------------------------------------------------
>>>>?The?NEW?KODAK?i700?Series?Scanners?deliver?under?ANY?circumstances!?Your
>>>>?production?scanning?environment?may?not?be?a?perfect?world?-?but?thanks?to
>>>>?Kodak,?there's?a?perfect?scanner?to?get?the?job?done!?With?the?NEW?KODAK
>>>>?i700
>>>>?Series?Scanner?you'll?get?full?speed?at?300?dpi?even?with?all?image
>>>>?processing?features?enabled.?http://p.sf.net/sfu/kodak-com
>>>>?_______________________________________________
>>>>?Ganglia-general?mailing?list
>>>>?ganglia-gene...@lists.sourceforge.net
>>>>?https://lists.sourceforge.net/lists/listinfo/ganglia-general
>>>>
>>>>
>>>
>>>Jerry,
>>>
>>>I?cannot?help?you?with?the?solution?but?I?can?help?collect?some
>>>information?so?other?don't?have?to?ask.
>>>
>>>What?is?your?ganglia?version?
>>>What?is?the?cluster?scale??(How?many?monitored?nodes?do?you?have?)
>>>How?long?has?the?cluster?been?up?
>>>And?what?is?the?OS?of?your?head?node?
>>>Does?it?always?stay?like?this?
>>>
>>>Please?kindly?provide?answers?to?the?question?above?and?I?hope?those
>>>will?be?helpful?for?investigation?by?others
>>>.
>>>--
>>>Regards,
>>>Simon?Yan
>>
>>
>>?________________________________
>>?穿越地震带?纪念汶川地震一周年
>>?------------------------------------------------------------------------------
>>?The?NEW?KODAK?i700?Series?Scanners?deliver?under?ANY?circumstances!?Your
>>?production?scanning?environment?may?not?be?a?perfect?world?-?but?thanks?to
>>?Kodak,?there's?a?perfect?scanner?to?get?the?job?done!?With?the?NEW?KODAK
>>?i700
>>?Series?Scanner?you'll?get?full?speed?at?300?dpi?even?with?all?image
>>?processing?features?enabled.?http://p.sf.net/sfu/kodak-com
>>?_______________________________________________
>>?Ganglia-general?mailing?list
>>?ganglia-gene...@lists.sourceforge.net
>>?https://lists.sourceforge.net/lists/listinfo/ganglia-general
>>
>>
>
>
>
>--?
>Jesse?Becker
>GPG?Fingerprint?--?BD00?7AA4?4483?AFCC?82D0??2720?0083?0931?9A2B?06A2
穿越地震带 纪念汶川地震一周年------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables
unlimited royalty-free distribution of the report engine
for externally facing server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers
