<?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: File table overflow. in Operating System - HP-UX</title>
    <link>https://community.hpe.com/t5/operating-system-hp-ux/file-table-overflow/m-p/3044676#M135052</link>
    <description>There are many things wrong with the system file you posted:&lt;BR /&gt;&lt;BR /&gt;maxdsiz is 3Gb which SAM may not like at all. Typically, SAM allows maxdsiz to be just about 2Gb. There are special requirements to allow a 32bit program just to get to 1.7Gb, and many others to go beyond 2Gb (including special patches). Change this to 2Gb for now.&lt;BR /&gt;&lt;BR /&gt;fs_async = 1 is a high rik parameter. Any unexpected power fail or a crash can leave busy filesystems corrupted so badly that fsck cannot fix them. fs_async=1 can improve write speed a bit, but the risk is too high for production systems. Also, fs_async=1 appears multiple times in the system file. Normally, SAM won't ever create a duplicate parameter entry and the extras may cause SAM to have difficulties.&lt;BR /&gt;&lt;BR /&gt;nfile and nproc don't make much sense. nproc is 7500 and since all processes need a minimum of 3 files (average is more like 5-6), then nfile should be 22500 to about 45000. Do you really have more than 5000 processes running at the same time? The rule of thumb is nfile = nproc * 5&lt;BR /&gt;&lt;BR /&gt;ninode = 8000 looks way too high. Change it to a fixed value, no more than 4096. &lt;BR /&gt;&lt;BR /&gt;maxusers = 128 yet you said you tried to change it maxusers from 32 to 64, yet the system file says 128. Has the system file been edited by hand? The inconsistencies seem unusual for SAM to allow.  To see the parameters your currently running kernel is using, use kmtune.&lt;BR /&gt;&lt;BR /&gt;max_thread_proc = 8000 which seems very high. DO you really have a process that will start more than a few hundred threads? A process that tries to run 7000+ threads needs a big design review.&lt;BR /&gt;-------&lt;BR /&gt;Compare the output of kmtune with SAM's list of parameters. They should be the same. Then shutdown all applications and lockout users until you can generate a new kernel. The file table overflow is caused by having too many users and processes on the system. Once a new kernel has been put into place with stable values, things should be a lot better.&lt;BR /&gt;&lt;BR /&gt;And most important, keep up to date on patches.</description>
    <pubDate>Sat, 09 Aug 2003 13:21:58 GMT</pubDate>
    <dc:creator>Bill Hassell</dc:creator>
    <dc:date>2003-08-09T13:21:58Z</dc:date>
    <item>
      <title>File table overflow.</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/file-table-overflow/m-p/3044673#M135049</link>
      <description>When i tried to change the value of Maxusers and nfile value in kernel configuration in SAM iam getting the following error.&lt;BR /&gt;&lt;BR /&gt;Cannot build a kernel using /usr/sbin/config.  The standard output of the command is shown below:&lt;BR /&gt;WARNING: No such device, tunable, or cdio autofsc&lt;BR /&gt;Ignoring the following line.&lt;BR /&gt;autofsc&lt;BR /&gt;(Bundled) cc: error 1404: Can't find the process "/usr/ccs/lbin/cpp". &lt;BR /&gt;stop.&lt;BR /&gt;config: make did an exit(1)&lt;BR /&gt;The command exitted with exitcode -71. &lt;BR /&gt;&lt;BR /&gt;Kindly help me out. Thanks in advance.&lt;BR /&gt;Regards&lt;BR /&gt;Jagadesh</description>
      <pubDate>Fri, 08 Aug 2003 03:31:10 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/file-table-overflow/m-p/3044673#M135049</guid>
      <dc:creator>Jagadesh</dc:creator>
      <dc:date>2003-08-08T03:31:10Z</dc:date>
    </item>
    <item>
      <title>Re: File table overflow.</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/file-table-overflow/m-p/3044674#M135050</link>
      <description>You could have a value that is outside the threshold if you increase some parameters too large. Try to lower the values and see if the kernel will build. Alternatively let us know what is in the system file and values your attempting to add. The error message however points to a problem with the 'autofsc' driver which should not be a problem. Have you also added 'cachefsc' ?</description>
      <pubDate>Fri, 08 Aug 2003 03:38:49 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/file-table-overflow/m-p/3044674#M135050</guid>
      <dc:creator>Michael Tully</dc:creator>
      <dc:date>2003-08-08T03:38:49Z</dc:date>
    </item>
    <item>
      <title>Re: File table overflow.</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/file-table-overflow/m-p/3044675#M135051</link>
      <description>Hi Michael,&lt;BR /&gt;&lt;BR /&gt;Pls have a look at my system file&lt;BR /&gt;&lt;BR /&gt;* Drivers and Subsystems&lt;BR /&gt;&lt;BR /&gt;DlkmDrv&lt;BR /&gt;GSCtoPCI&lt;BR /&gt;PCItoPCI&lt;BR /&gt;SCentIf&lt;BR /&gt;arp&lt;BR /&gt;asio0&lt;BR /&gt;asp&lt;BR /&gt;audio&lt;BR /&gt;autofsc&lt;BR /&gt;beep&lt;BR /&gt;btlan3&lt;BR /&gt;btlan6&lt;BR /&gt;c720&lt;BR /&gt;cb&lt;BR /&gt;cdfs&lt;BR /&gt;clone&lt;BR /&gt;core&lt;BR /&gt;diag1&lt;BR /&gt;diag2&lt;BR /&gt;dlkm&lt;BR /&gt;dlpi&lt;BR /&gt;dmem&lt;BR /&gt;echo&lt;BR /&gt;fc&lt;BR /&gt;fc_arp&lt;BR /&gt;fcgsc&lt;BR /&gt;fcgsc_lan&lt;BR /&gt;ffs&lt;BR /&gt;foreign&lt;BR /&gt;framebuf&lt;BR /&gt;graph3&lt;BR /&gt;hcd&lt;BR /&gt;hid&lt;BR /&gt;hpstreams&lt;BR /&gt;hub&lt;BR /&gt;inet&lt;BR /&gt;ip&lt;BR /&gt;ite&lt;BR /&gt;klog&lt;BR /&gt;lasi&lt;BR /&gt;lba&lt;BR /&gt;ldterm&lt;BR /&gt;lvmaclan&lt;BR /&gt;netdiag1&lt;BR /&gt;netqa&lt;BR /&gt;nfs_client&lt;BR /&gt;nfs_core&lt;BR /&gt;nfs_server&lt;BR /&gt;nfsm&lt;BR /&gt;nms&lt;BR /&gt;nuls&lt;BR /&gt;pa_generic_psm&lt;BR /&gt;pa_psm&lt;BR /&gt;pat_psm&lt;BR /&gt;pci&lt;BR /&gt;pckt&lt;BR /&gt;pipedev&lt;BR /&gt;pipemod&lt;BR /&gt;ptem&lt;BR /&gt;ptm&lt;BR /&gt;pts&lt;BR /&gt;rawip&lt;BR /&gt;rpcmod&lt;BR /&gt;sad&lt;BR /&gt;sapic&lt;BR /&gt;lvm&lt;BR /&gt;sba&lt;BR /&gt;sc&lt;BR /&gt;sctl&lt;BR /&gt;sdisk&lt;BR /&gt;side&lt;BR /&gt;stape&lt;BR /&gt;stcpmap&lt;BR /&gt;strlog&lt;BR /&gt;strpty_included&lt;BR /&gt;strtelnet_included&lt;BR /&gt;superio&lt;BR /&gt;tcp&lt;BR /&gt;telm&lt;BR /&gt;tels&lt;BR /&gt;timod&lt;BR /&gt;tirdwr&lt;BR /&gt;tlclts&lt;BR /&gt;tlcots&lt;BR /&gt;tlcotsod&lt;BR /&gt;token_arp&lt;BR /&gt;tun&lt;BR /&gt;udp&lt;BR /&gt;ufs&lt;BR /&gt;uipc&lt;BR /&gt;usbd&lt;BR /&gt;vxbase&lt;BR /&gt;wsio&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;* Kernel Device info&lt;BR /&gt;&lt;BR /&gt;dump lvol&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;* Tunable parameters&lt;BR /&gt;&lt;BR /&gt;STRMSGSZ        65535&lt;BR /&gt;create_fastlinks 1&lt;BR /&gt;default_disk_ir 1&lt;BR /&gt;fs_async        1&lt;BR /&gt;max_thread_proc 8000&lt;BR /&gt;maxdsiz         3221225472&lt;BR /&gt;maxdsiz_64bit   17179869184&lt;BR /&gt;maxfiles        2048&lt;BR /&gt;maxfiles_lim    2048&lt;BR /&gt;maxssiz         83570668&lt;BR /&gt;default_disk_ir 1&lt;BR /&gt;fs_async        1&lt;BR /&gt;max_thread_proc 8000&lt;BR /&gt;maxdsiz         3221225472&lt;BR /&gt;maxdsiz_64bit   17179869184&lt;BR /&gt;maxfiles        2048&lt;BR /&gt;maxfiles_lim    2048&lt;BR /&gt;maxssiz         83570668&lt;BR /&gt;maxssiz_64bit   268435456&lt;BR /&gt;maxswapchunks   4096&lt;BR /&gt;maxtsiz         0X40000000&lt;BR /&gt;maxtsiz_64bit   4294967296&lt;BR /&gt;maxuprc         256&lt;BR /&gt;maxusers        128&lt;BR /&gt;ncsize          5024&lt;BR /&gt;nfile           8000&lt;BR /&gt;ninode          8000&lt;BR /&gt;nkthread        15000&lt;BR /&gt;nproc           7500&lt;BR /&gt;npty            200&lt;BR /&gt;nstrpty         200&lt;BR /&gt;shmmax          750741824&lt;BR /&gt;dbc_max_pct 20&lt;BR /&gt;&lt;BR /&gt;I have tried to change the value of maxusers and nfile&lt;BR /&gt;maxusers from 32 to 64&lt;BR /&gt;nfile from 910 to 1981&lt;BR /&gt;but it is not allowing me to do the same. How can i overcome this file table overflow. Some times it is not allowing me to open the SAM itshelf. Iam getting&lt;BR /&gt;# /usr/sbin/sam&lt;BR /&gt;crt0: ERROR couldn't open /usr/lib/dld.sl errno:000000023&lt;BR /&gt;/usr/sbin/sam[196]: /var/sam: bad directory&lt;BR /&gt;/usr/sbin/sam[222]: rsam:  not found&lt;BR /&gt;&lt;BR /&gt;Kindly help me out&lt;BR /&gt;&lt;BR /&gt;</description>
      <pubDate>Sat, 09 Aug 2003 09:25:53 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/file-table-overflow/m-p/3044675#M135051</guid>
      <dc:creator>Jagadesh</dc:creator>
      <dc:date>2003-08-09T09:25:53Z</dc:date>
    </item>
    <item>
      <title>Re: File table overflow.</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/file-table-overflow/m-p/3044676#M135052</link>
      <description>There are many things wrong with the system file you posted:&lt;BR /&gt;&lt;BR /&gt;maxdsiz is 3Gb which SAM may not like at all. Typically, SAM allows maxdsiz to be just about 2Gb. There are special requirements to allow a 32bit program just to get to 1.7Gb, and many others to go beyond 2Gb (including special patches). Change this to 2Gb for now.&lt;BR /&gt;&lt;BR /&gt;fs_async = 1 is a high rik parameter. Any unexpected power fail or a crash can leave busy filesystems corrupted so badly that fsck cannot fix them. fs_async=1 can improve write speed a bit, but the risk is too high for production systems. Also, fs_async=1 appears multiple times in the system file. Normally, SAM won't ever create a duplicate parameter entry and the extras may cause SAM to have difficulties.&lt;BR /&gt;&lt;BR /&gt;nfile and nproc don't make much sense. nproc is 7500 and since all processes need a minimum of 3 files (average is more like 5-6), then nfile should be 22500 to about 45000. Do you really have more than 5000 processes running at the same time? The rule of thumb is nfile = nproc * 5&lt;BR /&gt;&lt;BR /&gt;ninode = 8000 looks way too high. Change it to a fixed value, no more than 4096. &lt;BR /&gt;&lt;BR /&gt;maxusers = 128 yet you said you tried to change it maxusers from 32 to 64, yet the system file says 128. Has the system file been edited by hand? The inconsistencies seem unusual for SAM to allow.  To see the parameters your currently running kernel is using, use kmtune.&lt;BR /&gt;&lt;BR /&gt;max_thread_proc = 8000 which seems very high. DO you really have a process that will start more than a few hundred threads? A process that tries to run 7000+ threads needs a big design review.&lt;BR /&gt;-------&lt;BR /&gt;Compare the output of kmtune with SAM's list of parameters. They should be the same. Then shutdown all applications and lockout users until you can generate a new kernel. The file table overflow is caused by having too many users and processes on the system. Once a new kernel has been put into place with stable values, things should be a lot better.&lt;BR /&gt;&lt;BR /&gt;And most important, keep up to date on patches.</description>
      <pubDate>Sat, 09 Aug 2003 13:21:58 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/file-table-overflow/m-p/3044676#M135052</guid>
      <dc:creator>Bill Hassell</dc:creator>
      <dc:date>2003-08-09T13:21:58Z</dc:date>
    </item>
  </channel>
</rss>

