Re: [expert] Fwd: memory leak
On Sun, 11 Aug 2002 18:32:33 -0700 gene [EMAIL PROTECTED] scribbled intuitively: snippage OK, here are the results of free and top. I can see that a lot of memory is cached. So you're saying that the cached memory is actually available? free -m total used free sharedbuffers cached Mem: 753750 2 0 11504 -/+ buffers/cache:234518 Swap: 784 0784 top 6:29pm up 23:15, 2 users, load average: 0.06, 0.18, 0.15 92 processes: 91 sleeping, 1 running, 0 zombie, 0 stopped CPU states: 6.2% user, 30.5% system, 0.0% nice, 63.1% idle Mem: 771192K av, 768136K used,3056K free, 116K shrd, 11384K buff Swap: 803176K av, 0K used, 803176K free 514912K cached PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND 26307 apache 9 0 14388 14M 2872 S 0.0 1.8 0:04 httpd-perl 26302 apache 9 0 14228 13M 2876 S 0.0 1.8 0:04 httpd-perl 26301 apache 9 0 14092 13M 2876 S 0.0 1.8 0:04 httpd-perl 26303 apache 9 0 14032 13M 2872 S 0.0 1.8 0:03 httpd-perl 26306 apache 9 0 14024 13M 2876 S 0.0 1.8 0:03 httpd-perl 26319 apache 9 0 13964 13M 2868 S 0.0 1.8 0:02 httpd-perl 26305 apache 9 0 13848 13M 2872 S 0.7 1.7 0:02 httpd-perl 26311 apache 9 0 13316 13M 2876 S 0.0 1.7 0:02 httpd-perl 26312 apache 9 0 13120 12M 2868 S 0.0 1.7 0:03 httpd-perl 2081 root 12 0 12980 12M 1848 S 5.0 1.6 28:25 server.pl 26318 apache 9 0 12836 12M 2872 S 0.0 1.6 0:02 httpd-perl 2240 mysql 9 0 5136 5136 2168 S 0.0 0.6 0:00 mysqld 2242 mysql 8 0 5136 5136 2168 S 0.0 0.6 0:00 mysqld 2243 mysql 9 0 5136 5136 2168 S 0.0 0.6 0:00 mysqld 2244 mysql 9 0 5136 5136 2168 S 0.0 0.6 0:00 mysqld 1443 root 9 0 3888 3888 1592 S 0.0 0.5 0:00 ntpd 1447 root 9 0 3888 3888 1592 S 0.0 0.5 0:00 ntpd 1452 root 9 0 3888 3888 1592 S 0.0 0.5 0:00 ntpd 1391 snort 9 0 3624 3624 1140 S 0.3 0.4 0:25 snort 1753 xfs9 0 3448 3448 912 S 0.0 0.4 0:00 xfs 26270 root 8 0 3180 3180 2456 S 0.0 0.4 0:00 httpd-perl 1415 named 9 0 2912 2912 1768 S 0.0 0.3 0:00 named 1419 named 9 0 2912 2912 1768 S 0.0 0.3 0:00 named 1420 named 9 0 2912 2912 1768 S 0.0 0.3 0:08 named 1425 named 9 0 2912 2912 1768 S 0.0 0.3 0:00 named 1426 named 9 0 2912 2912 1768 S 0.0 0.3 0:01 named 26740 gene 9 0 2104 2104 1888 S 0.7 0.2 0:00 sshd 25221 gene 9 0 2036 2036 1652 S 0.0 0.2 0:00 imapsd 25222 admin 9 0 2012 2012 1648 S 0.0 0.2 0:00 imapsd 26738 root 9 0 1912 1912 1720 S 0.0 0.2 0:00 sshd 26299 apache 9 0 1788 1788 1488 S 0.0 0.2 0:00 httpd 26300 apache 9 0 1788 1788 1488 S 0.0 0.2 0:00 httpd 26304 apache 9 0 1784 1784 1488 S 0.0 0.2 0:00 httpd 26282 apache 9 0 1756 1756 1484 S 0.0 0.2 0:00 httpd 26308 apache 9 0 1748 1748 1484 S 0.0 0.2 0:00 httpd 26283 apache 9 0 1744 1744 1484 S 0.0 0.2 0:00 httpd 26293 apache 9 0 1744 1744 1484 S 0.0 0.2 0:00 httpd 26294 apache 9 0 1744 1744 1484 S 0.0 0.2 0:00 httpd 26295 apache 9 0 1744 1744 1484 S 0.0 0.2 0:00 httpd 26298 apache 9 0 1744 1744 1484 S 0.0 0.2 0:00 httpd 2131 root 8 0 1644 1644 1112 S 0.0 0.2 0:01 bash 26741 gene 9 0 1576 1576 1088 S 0.0 0.2 0:00 bash 26281 root 8 0 1448 1448 1272 S 0.0 0.1 0:00 httpd == Looks as if the bulk of your mem (over half a gig) is simply being cached by the system. I'm not sure that's a problem as I believe it will be released when need for apps or processes. Others, more knowledgeable might have better info for you. Mike -- Give me a woman who loves beer and I will conquer the world. -- Kaiser Wilhelm Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] Fwd: memory leak
Haven't tried it yet but just found it on my install disks. The rpm is insure++5.2-1mdk and it's supposed to aimed right at memory leaks in C/C++ programs. James On Sat, 10 Aug 2002 18:43:12 -0700 gene [EMAIL PROTECTED] Grabed a keyboard and said: Does anyone have advice for tracking down a memory leak? I'm getting down to 2M of RAM. Running 'top' doesn't show anything using much memory, and I've tried killing most of the running processes but that doesn't help. I can only free up the memory by restarting my server. I'm running kernel 2.4.7-12.3mdk. Any advice is appreciated. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] Fwd: memory leak
On Saturday, August 10, 2002, at 07:01 PM, Michael Scottaline wrote: On Sat, 10 Aug 2002 18:43:12 -0700 gene [EMAIL PROTECTED] scribbled intuitively: Does anyone have advice for tracking down a memory leak? I'm getting down to 2M of RAM. Running 'top' doesn't show anything using much memory, and I've tried killing most of the running processes but that doesn't help. I can only free up the memory by restarting my server. I'm running kernel 2.4.7-12.3mdk. Any advice is appreciated. What *exactly* is top reporting to you. You may not actually have a memory leak. Linux tends to view unused memory as wasted memory. Howmuch of that mem is actually cached. OK, here are the results of free and top. I can see that a lot of memory is cached. So you're saying that the cached memory is actually available? free -m total used free sharedbuffers cached Mem: 753750 2 0 11504 -/+ buffers/cache:234518 Swap: 784 0784 top 6:29pm up 23:15, 2 users, load average: 0.06, 0.18, 0.15 92 processes: 91 sleeping, 1 running, 0 zombie, 0 stopped CPU states: 6.2% user, 30.5% system, 0.0% nice, 63.1% idle Mem: 771192K av, 768136K used,3056K free, 116K shrd, 11384K buff Swap: 803176K av, 0K used, 803176K free 514912K cached PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND 26307 apache 9 0 14388 14M 2872 S 0.0 1.8 0:04 httpd-perl 26302 apache 9 0 14228 13M 2876 S 0.0 1.8 0:04 httpd-perl 26301 apache 9 0 14092 13M 2876 S 0.0 1.8 0:04 httpd-perl 26303 apache 9 0 14032 13M 2872 S 0.0 1.8 0:03 httpd-perl 26306 apache 9 0 14024 13M 2876 S 0.0 1.8 0:03 httpd-perl 26319 apache 9 0 13964 13M 2868 S 0.0 1.8 0:02 httpd-perl 26305 apache 9 0 13848 13M 2872 S 0.7 1.7 0:02 httpd-perl 26311 apache 9 0 13316 13M 2876 S 0.0 1.7 0:02 httpd-perl 26312 apache 9 0 13120 12M 2868 S 0.0 1.7 0:03 httpd-perl 2081 root 12 0 12980 12M 1848 S 5.0 1.6 28:25 server.pl 26318 apache 9 0 12836 12M 2872 S 0.0 1.6 0:02 httpd-perl 2240 mysql 9 0 5136 5136 2168 S 0.0 0.6 0:00 mysqld 2242 mysql 8 0 5136 5136 2168 S 0.0 0.6 0:00 mysqld 2243 mysql 9 0 5136 5136 2168 S 0.0 0.6 0:00 mysqld 2244 mysql 9 0 5136 5136 2168 S 0.0 0.6 0:00 mysqld 1443 root 9 0 3888 3888 1592 S 0.0 0.5 0:00 ntpd 1447 root 9 0 3888 3888 1592 S 0.0 0.5 0:00 ntpd 1452 root 9 0 3888 3888 1592 S 0.0 0.5 0:00 ntpd 1391 snort 9 0 3624 3624 1140 S 0.3 0.4 0:25 snort 1753 xfs9 0 3448 3448 912 S 0.0 0.4 0:00 xfs 26270 root 8 0 3180 3180 2456 S 0.0 0.4 0:00 httpd-perl 1415 named 9 0 2912 2912 1768 S 0.0 0.3 0:00 named 1419 named 9 0 2912 2912 1768 S 0.0 0.3 0:00 named 1420 named 9 0 2912 2912 1768 S 0.0 0.3 0:08 named 1425 named 9 0 2912 2912 1768 S 0.0 0.3 0:00 named 1426 named 9 0 2912 2912 1768 S 0.0 0.3 0:01 named 26740 gene 9 0 2104 2104 1888 S 0.7 0.2 0:00 sshd 25221 gene 9 0 2036 2036 1652 S 0.0 0.2 0:00 imapsd 25222 admin 9 0 2012 2012 1648 S 0.0 0.2 0:00 imapsd 26738 root 9 0 1912 1912 1720 S 0.0 0.2 0:00 sshd 26299 apache 9 0 1788 1788 1488 S 0.0 0.2 0:00 httpd 26300 apache 9 0 1788 1788 1488 S 0.0 0.2 0:00 httpd 26304 apache 9 0 1784 1784 1488 S 0.0 0.2 0:00 httpd 26282 apache 9 0 1756 1756 1484 S 0.0 0.2 0:00 httpd 26308 apache 9 0 1748 1748 1484 S 0.0 0.2 0:00 httpd 26283 apache 9 0 1744 1744 1484 S 0.0 0.2 0:00 httpd 26293 apache 9 0 1744 1744 1484 S 0.0 0.2 0:00 httpd 26294 apache 9 0 1744 1744 1484 S 0.0 0.2 0:00 httpd 26295 apache 9 0 1744 1744 1484 S 0.0 0.2 0:00 httpd 26298 apache 9 0 1744 1744 1484 S 0.0 0.2 0:00 httpd 2131 root 8 0 1644 1644 1112 S 0.0 0.2 0:01 bash 26741 gene 9 0 1576 1576 1088 S 0.0 0.2 0:00 bash 26281 root 8 0 1448 1448 1272 S 0.0 0.1 0:00 httpd Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] Fwd: memory leak
total used free sharedbuffers cached Mem: 753750 2 0 11504 -/+ buffers/cache:234518 Swap: 784 0784 Well, that seems reasonable. Cached RAM is pretty high, but that just depends on prior activity - for instance using gcc a lot or other big jobs, the cache will get utilized to remember those pages because they may be pressed into service again (like another gcc run). But if processes come along and request RAM, the OS will just reuse what's in buffers and cache. So, your memory isn't 'gone' it is just allocated in case something comes along again that needs it. Buffers (i.e., disk cache) get reused for process memory if there isn't disk activity going on -- if there's plenty of memory available elsewhere, buffers tend to not take up that much space, unless there's a lot of recent disk activity. For instance if you get up in the morning you may notice a large number in buffers first thing in the morning. That's because 'locate' was just run a few hours before -- and that works the disk. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] Fwd: memory leak
On Sat, 10 Aug 2002 18:43:12 -0700 gene [EMAIL PROTECTED] scribbled intuitively: Does anyone have advice for tracking down a memory leak? I'm getting down to 2M of RAM. Running 'top' doesn't show anything using much memory, and I've tried killing most of the running processes but that doesn't help. I can only free up the memory by restarting my server. I'm running kernel 2.4.7-12.3mdk. Any advice is appreciated. What *exactly* is top reporting to you. You may not actually have a memory leak. Linux tends to view unused memory as wasted memory. Howmuch of that mem is actually cached. Mike -- Education is a progressive discovery of our own ignorance. --Will Durant Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
