<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Re: Out of memory in Operating System - Linux</title>
    <link>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473799#M37791</link>
    <description>The rule about looking at "free" output and adding the free and cache&lt;BR /&gt;values is generally correct.  But it oversimplifies.  Much of the&lt;BR /&gt;cache memory is easily released for reuse.  But shared memory segments&lt;BR /&gt;can appear as "cached" memory although that cannot actually be freed.&lt;BR /&gt;If you look at the output of "ipcs -u" you can see the size of shared&lt;BR /&gt;memory segments.  The values are reported in terms of pages rather&lt;BR /&gt;than 1K or 1M.  The page size is almost always 4K, but varies on some&lt;BR /&gt;architectures.  You can check that with the "getconf PAGESIZE" command.&lt;BR /&gt;Shared memory can be pushed from RAM to swap if swap space is available.&lt;BR /&gt;Doing that will be slower than releasing clean cache memory, since the&lt;BR /&gt;data will need to be written out to disk.  If swap space runs out then&lt;BR /&gt;a system can start doing out of memory kills even though there is still&lt;BR /&gt;a lot of "cached" memory remaining.&lt;BR /&gt;</description>
    <pubDate>Thu, 06 Aug 2009 16:40:34 GMT</pubDate>
    <dc:creator>Mike Stroyan</dc:creator>
    <dc:date>2009-08-06T16:40:34Z</dc:date>
    <item>
      <title>Out of memory</title>
      <link>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473794#M37786</link>
      <description>Dear all,&lt;BR /&gt;&lt;BR /&gt;I've got 2 issues:&lt;BR /&gt;&lt;BR /&gt;1. a RHEL Update 4 node was hanging with the error " Out Of Memory: Killed Process 21344 ( oracle ). It's a X64. What could be the issue?&lt;BR /&gt;&lt;BR /&gt;2. Well, what is the most accurate way to get the free memory of the server. I was told to check:&lt;BR /&gt;&lt;BR /&gt;free -m&lt;BR /&gt;             total       used       free     shared    buffers     cached&lt;BR /&gt;Mem:         16010      15990         20          0         69      10615&lt;BR /&gt;-/+ buffers/cache:       5304      10705&lt;BR /&gt;Swap:         4031       2314       1717&lt;BR /&gt;&lt;BR /&gt;according to the above command, the system has 10705 free memory.&lt;BR /&gt;&lt;BR /&gt;- top says 21180k free&lt;BR /&gt;&lt;BR /&gt;- vmstat says 21300 free.&lt;BR /&gt;&lt;BR /&gt;- meminfo says MemFree:         42556 kB&lt;BR /&gt;&lt;BR /&gt;But there is a minute time diff as i was copying and pasting across. Please enlighten me.&lt;BR /&gt;&lt;BR /&gt;cheers&lt;BR /&gt;</description>
      <pubDate>Wed, 05 Aug 2009 11:34:28 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473794#M37786</guid>
      <dc:creator>ossupport55</dc:creator>
      <dc:date>2009-08-05T11:34:28Z</dc:date>
    </item>
    <item>
      <title>Re: Out of memory</title>
      <link>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473795#M37787</link>
      <description>Hi,&lt;BR /&gt;&lt;BR /&gt;How much memory there in your server???&lt;BR /&gt;Check how much of memory the oracle are use.&lt;BR /&gt;from the output of all commands, there is not enough memory to run normal system ( almost 42MB free), so I think you have bad memory utilization or you really need more memory for your system.</description>
      <pubDate>Wed, 05 Aug 2009 12:40:57 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473795#M37787</guid>
      <dc:creator>Fayez</dc:creator>
      <dc:date>2009-08-05T12:40:57Z</dc:date>
    </item>
    <item>
      <title>Re: Out of memory</title>
      <link>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473796#M37788</link>
      <description>The real used memory for applications it's what you see on +/- buffers/cache. This has been discussed a lot in this forum.&lt;BR /&gt;&lt;BR /&gt;It's impossible to determine the cause of the problem with the information provided. I suggest you to install "collect for linux" collectl and capture performance information.&lt;BR /&gt;&lt;BR /&gt;</description>
      <pubDate>Wed, 05 Aug 2009 12:45:10 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473796#M37788</guid>
      <dc:creator>Ivan Ferreira</dc:creator>
      <dc:date>2009-08-05T12:45:10Z</dc:date>
    </item>
    <item>
      <title>Re: Out of memory</title>
      <link>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473797#M37789</link>
      <description>the oom killer kicked in. Please provide this output:&lt;BR /&gt;&lt;BR /&gt;sysctl vm.overcommit_memory</description>
      <pubDate>Wed, 05 Aug 2009 12:56:00 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473797#M37789</guid>
      <dc:creator>Court Campbell</dc:creator>
      <dc:date>2009-08-05T12:56:00Z</dc:date>
    </item>
    <item>
      <title>Re: Out of memory</title>
      <link>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473798#M37790</link>
      <description>Thanks to all,&lt;BR /&gt;&lt;BR /&gt;Court Campbell, yes out of memory setting is at it's default.&lt;BR /&gt;vm.overcommit_memory = 0&lt;BR /&gt;should i make it 2. i think if it's to be done it should be done on a reactive basis rather than proactive. We have over 300 servers.&lt;BR /&gt;Ivan, can you please provide me some reference about memory usage and free memory allocation. According to my post, vmstat,top and the meminfo file is useless is it ?&lt;BR /&gt;Fayez, MemTotal:     16394912 kB is the one which is reported in /proc/meminfo, well now i have doubts whether these are really acceptable readings.&lt;BR /&gt;&lt;BR /&gt;Finally, gurus please provide me some reference on how to get to know the actual memory utilisation physical/virtual.&lt;BR /&gt;&lt;BR /&gt;cheers!</description>
      <pubDate>Thu, 06 Aug 2009 03:53:56 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473798#M37790</guid>
      <dc:creator>ossupport55</dc:creator>
      <dc:date>2009-08-06T03:53:56Z</dc:date>
    </item>
    <item>
      <title>Re: Out of memory</title>
      <link>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473799#M37791</link>
      <description>The rule about looking at "free" output and adding the free and cache&lt;BR /&gt;values is generally correct.  But it oversimplifies.  Much of the&lt;BR /&gt;cache memory is easily released for reuse.  But shared memory segments&lt;BR /&gt;can appear as "cached" memory although that cannot actually be freed.&lt;BR /&gt;If you look at the output of "ipcs -u" you can see the size of shared&lt;BR /&gt;memory segments.  The values are reported in terms of pages rather&lt;BR /&gt;than 1K or 1M.  The page size is almost always 4K, but varies on some&lt;BR /&gt;architectures.  You can check that with the "getconf PAGESIZE" command.&lt;BR /&gt;Shared memory can be pushed from RAM to swap if swap space is available.&lt;BR /&gt;Doing that will be slower than releasing clean cache memory, since the&lt;BR /&gt;data will need to be written out to disk.  If swap space runs out then&lt;BR /&gt;a system can start doing out of memory kills even though there is still&lt;BR /&gt;a lot of "cached" memory remaining.&lt;BR /&gt;</description>
      <pubDate>Thu, 06 Aug 2009 16:40:34 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-linux/out-of-memory/m-p/4473799#M37791</guid>
      <dc:creator>Mike Stroyan</dc:creator>
      <dc:date>2009-08-06T16:40:34Z</dc:date>
    </item>
  </channel>
</rss>

