<?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: XFC and memory management in Operating System - OpenVMS</title>
    <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196752#M89761</link>
    <description>And the min size is missing in 7.3.</description>
    <pubDate>Tue, 13 May 2008 13:18:27 GMT</pubDate>
    <dc:creator>Wim Van den Wyngaert</dc:creator>
    <dc:date>2008-05-13T13:18:27Z</dc:date>
    <item>
      <title>XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196742#M89751</link>
      <description>The immediate environment is Alpha OpenVMS v7.3-2 but if I’d also be interested to know if behavior changes in subsequent VMS versions. The system in question was configured to use XFC with VCC_MAX_CACHE set to -1. Over time the XFC had become fully populated, consuming 50% of physical memory (4GB out of 8GB). With the typical full complement of users this system would normally never page or swap as there was expected to be adequate physical memory to support large working set extents for the expected number of processes. However, on one recent occasion, a large number of additional processes were created. One side effect of this was swapping. I was surprised by this – rather, I expected that the XFC would begin relinquishing some of that 50% of physical memory that it was consuming before swapping occurred – it didn’t and the swap file space was inadequate (this has been remedied – a secondary swap file was accidentally lost during a storage upgrade when several old small disks were consolidated into one larger one). Is this expected behavior or should swapping have been postponed until the XFC had surrendered some of what it was holding? Thanks for reading this and comments that you might have – more info can be provided if required.</description>
      <pubDate>Tue, 13 May 2008 11:22:39 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196742#M89751</guid>
      <dc:creator>Jim_McKinney</dc:creator>
      <dc:date>2008-05-13T11:22:39Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196743#M89752</link>
      <description>Of an old thread :&lt;BR /&gt;&lt;BR /&gt;Tested it on 7.3.&lt;BR /&gt;&lt;BR /&gt;Workstation with 256 MB. VCC_MAX on -1 thus 128 MB is the max size.&lt;BR /&gt;&lt;BR /&gt;Allocated 200 MB and made it dirty.&lt;BR /&gt;&lt;BR /&gt;Again and again. At the end the pagefile was full but XFC still used 12 MB allocated(started the test with 77 MB).&lt;BR /&gt;&lt;BR /&gt;BTW : I ran the malloc again for 5 MB and the system went into hang.&lt;BR /&gt;&lt;BR /&gt;Wim</description>
      <pubDate>Tue, 13 May 2008 11:31:23 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196743#M89752</guid>
      <dc:creator>Wim Van den Wyngaert</dc:creator>
      <dc:date>2008-05-13T11:31:23Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196744#M89753</link>
      <description>Hello&lt;BR /&gt;&lt;BR /&gt;I guess you know there is the sda extension xfc&lt;BR /&gt;$ ana/sys&lt;BR /&gt;xfc&lt;BR /&gt;help&lt;BR /&gt;&lt;BR /&gt;and &lt;BR /&gt;xfc sh file/brief&lt;BR /&gt;xfc show memory(/verify)&lt;BR /&gt;xfc sh summary/stat&lt;BR /&gt;&lt;BR /&gt;and similar commands regularly may help monitor this.&lt;BR /&gt;&lt;BR /&gt;This may vary with the patches applied on your system.&lt;BR /&gt;&lt;BR /&gt;Regards&lt;BR /&gt;&lt;BR /&gt;Gerard</description>
      <pubDate>Tue, 13 May 2008 11:40:02 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196744#M89753</guid>
      <dc:creator>labadie_1</dc:creator>
      <dc:date>2008-05-13T11:40:02Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196745#M89754</link>
      <description>Perhaps I should phrase this another way - if there's a run on the (memory free list) bank, how is it expected that the bank will be funded? - by trimming of the XFC or by outswapping idle processes? Who yields first? I'm interested in the ruleset that is applied here.</description>
      <pubDate>Tue, 13 May 2008 11:47:13 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196745#M89754</guid>
      <dc:creator>Jim_McKinney</dc:creator>
      <dc:date>2008-05-13T11:47:13Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196746#M89755</link>
      <description>To be complete : I used XFC 2.0 dd aug-2002. The last version is of JUL-2004.&lt;BR /&gt;&lt;BR /&gt;Wim</description>
      <pubDate>Tue, 13 May 2008 11:48:29 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196746#M89755</guid>
      <dc:creator>Wim Van den Wyngaert</dc:creator>
      <dc:date>2008-05-13T11:48:29Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196747#M89756</link>
      <description>&lt;!--!*#--&gt;$ pipe prod show history | sear sys$pipe xfc&lt;BR /&gt;DEC AXPVMS VMS732_XFC V4.0          Patch       Install     10-SEP-2007 20:19:07&lt;BR /&gt;&lt;BR /&gt;$ ana/sys&lt;BR /&gt;&lt;BR /&gt;OpenVMS (TM) system analyzer&lt;BR /&gt;&lt;BR /&gt;SDA&amp;gt; xfc show sum&lt;BR /&gt;&lt;BR /&gt;XFC Summary (current time 13-MAY-2008 05:55:37.58)&lt;BR /&gt;--------------------------------------------------&lt;BR /&gt;Extended File Cache V1.0 (Mar  8 2007 11:18:39)</description>
      <pubDate>Tue, 13 May 2008 11:56:32 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196747#M89756</guid>
      <dc:creator>Jim_McKinney</dc:creator>
      <dc:date>2008-05-13T11:56:32Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196748#M89757</link>
      <description>There is a patch XFC V4, have you applied it ?&lt;BR /&gt;&lt;BR /&gt;The releases notes say, among other things&lt;BR /&gt;&lt;BR /&gt;     &lt;BR /&gt;The most recent patch is :  VMS732_XFC-V0400&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;     5.2  Problems addressed in this kit&lt;BR /&gt;&lt;BR /&gt;          5.2.1  Limit the Maximum Number of Closed Files Retained in&lt;BR /&gt;                 Cache&lt;BR /&gt;&lt;BR /&gt;               5.2.1.1  Problem Description:&lt;BR /&gt;&lt;BR /&gt;               Excessive use of non-paged pool by XFC closed files could&lt;BR /&gt;               lead to non-paged pool allocation failures and bugchecks.&lt;BR /&gt;&lt;BR /&gt;               The maximum number of closed files that XFC will retain&lt;BR /&gt;               in the cache is now limited so that XFC will use no more&lt;BR /&gt;               than one half of the original size of non-paged pool (the&lt;BR /&gt;               SYSGEN parameter NPAGEDYN).  This value will be&lt;BR /&gt;               approximately 1500 files per megabyte of non-paged pool.&lt;BR /&gt;               For example, if the SYSGEN parameter NPAGEDYN is&lt;BR /&gt;               30,000,000 then then XFC will retain no more than about&lt;BR /&gt;               45,000 closed files in the cache.&lt;BR /&gt;</description>
      <pubDate>Tue, 13 May 2008 12:17:13 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196748#M89757</guid>
      <dc:creator>labadie_1</dc:creator>
      <dc:date>2008-05-13T12:17:13Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196749#M89758</link>
      <description>Did some reading and read that the reclaiming is somewhat slow. So may be it's a question of waiting ...&lt;BR /&gt;&lt;BR /&gt;Also : in VMS 7.3 doc I read&lt;BR /&gt;&lt;BR /&gt;$ RUN SYS$SYSTEM:SYSMAN&lt;BR /&gt;SYSMAN&amp;gt; RESERVED_MEMORY SHOW VCC$MIN_CACHE_SIZE /NOGLOBAL_SECTION&lt;BR /&gt;&lt;BR /&gt;should return the minimum size of the VCC. But not on my 7.3 system.&lt;BR /&gt;&lt;BR /&gt;Wim &lt;BR /&gt;</description>
      <pubDate>Tue, 13 May 2008 12:58:49 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196749#M89758</guid>
      <dc:creator>Wim Van den Wyngaert</dc:creator>
      <dc:date>2008-05-13T12:58:49Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196750#M89759</link>
      <description>Mmm you have to add the reserved memory yourself. I didn't.&lt;BR /&gt;&lt;BR /&gt;Wim</description>
      <pubDate>Tue, 13 May 2008 13:08:05 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196750#M89759</guid>
      <dc:creator>Wim Van den Wyngaert</dc:creator>
      <dc:date>2008-05-13T13:08:05Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196751#M89760</link>
      <description>With VMS 7.3-2 the SDA XFC extension will permit you to see the minimum cache size... a convenient system shows the following.&lt;BR /&gt;&lt;BR /&gt;SDA&amp;gt; XFC SHOW MEM&lt;BR /&gt;[ . . . ]&lt;BR /&gt;Boottime Maximum Cache Size      :    402653184 ( 384.0 MB)&lt;BR /&gt;Current Maximum Cache Size       :    402653184 ( 384.0 MB)&lt;BR /&gt;Total Allocated Cache Memory     :    396904832 ( 378.5 MB)  Peak:    397094784 ( 378.7 MB)&lt;BR /&gt;Minimum cache size               :      8052736 (   7.7 MB)&lt;BR /&gt;[ . . . ]</description>
      <pubDate>Tue, 13 May 2008 13:10:36 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196751#M89760</guid>
      <dc:creator>Jim_McKinney</dc:creator>
      <dc:date>2008-05-13T13:10:36Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196752#M89761</link>
      <description>And the min size is missing in 7.3.</description>
      <pubDate>Tue, 13 May 2008 13:18:27 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196752#M89761</guid>
      <dc:creator>Wim Van den Wyngaert</dc:creator>
      <dc:date>2008-05-13T13:18:27Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196753#M89762</link>
      <description>Did the same test again. Started with 58 MB and went down to 9.3 MB. And hang. Killed a decw process, waited a few minutes and the system came out of hang. 9.31 MB allocated.&lt;BR /&gt;&lt;BR /&gt;Fwiw&lt;BR /&gt;&lt;BR /&gt;Wim</description>
      <pubDate>Tue, 13 May 2008 13:44:40 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196753#M89762</guid>
      <dc:creator>Wim Van den Wyngaert</dc:creator>
      <dc:date>2008-05-13T13:44:40Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196754#M89763</link>
      <description>And as soon as I went out of phusical memory the XFC cache was decreased to 9.3 MB. So that's ok. Wonder why it wasn't decreased in your case.&lt;BR /&gt;&lt;BR /&gt;BTW : did you mean paging or swapping ?&lt;BR /&gt;&lt;BR /&gt;Wim</description>
      <pubDate>Tue, 13 May 2008 13:49:16 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196754#M89763</guid>
      <dc:creator>Wim Van den Wyngaert</dc:creator>
      <dc:date>2008-05-13T13:49:16Z</dc:date>
    </item>
    <item>
      <title>Re: XFC and memory management</title>
      <link>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196755#M89764</link>
      <description>I meant swapping - but paging to the page file was occurring as well. Processes were being trimmed back to their WSQUOTAs but the XFC wouldn't open its fingers and let go any of the memory that it had - at least not quickly enough to save this system.</description>
      <pubDate>Tue, 13 May 2008 14:15:25 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-openvms/xfc-and-memory-management/m-p/4196755#M89764</guid>
      <dc:creator>Jim_McKinney</dc:creator>
      <dc:date>2008-05-13T14:15:25Z</dc:date>
    </item>
  </channel>
</rss>

