LustreMonitoring.xml

   1 <?xml version="1.0" encoding="UTF-8"?>
   2 <chapter version="5.0" xml:lang="en-US" xmlns="http://docbook.org/ns/docbook" xmlns:xl="http://www.w3.org/1999/xlink" xml:id='lustremonitoring'>
   3   <info>
   4     <title xml:id='lustremonitoring.title'>Lustre Monitoring</title>
   5   </info>
   6   <para><anchor xml:id="dbdoclet.50438273_pgfId-1291311" xreflabel=""/>This chapter provides information on monitoring Lustre and includes the following sections:</para>
   7
   8   <itemizedlist><listitem>
   9           <para><xref linkend='dbdoclet.50438273_18711'/>Lustre Changelogs</para>
  10       </listitem>
  11       <listitem>
  12           <para><xref linkend="dbdoclet.50438273_81684"/>Lustre Monitoring Tool</para>
  13       </listitem>
  14       <listitem>
  15           <para><xref linkend="dbdoclet.50438273_80593"/>CollectL</para>
  16       </listitem>
  17       <listitem>
  18           <para><xref linkend="dbdoclet.50438273_44185"/>Other Monitoring Options</para>
  19           </listitem>
  20       </itemizedlist>
  21
  22     <section xml:id="dbdoclet.50438273_18711">
  23       <title>12.1 Lustre <anchor xml:id="dbdoclet.50438273_marker-1297383" xreflabel=""/>Changelogs</title>
  24       <para><anchor xml:id="dbdoclet.50438273_pgfId-1296794" xreflabel=""/>The changelogs feature records events that change the file system namespace or file metadata. Changes such as file creation, deletion, renaming, attribute changes, etc. are recorded with the target and parent file identifiers (FIDs), the name of the target, and a timestamp. These records can be used for a variety of purposes:</para>
  25       <itemizedlist><listitem>
  26           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296795" xreflabel=""/> Capture recent changes to feed into an archiving system.</para>
  27         </listitem>
  28 <listitem>
  29           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296796" xreflabel=""/> Use changelog entries to exactly replicate changes in a file system mirror.</para>
  30         </listitem>
  31 <listitem>
  32           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296797" xreflabel=""/> Set up &quot;watch scripts&quot; that take action on certain events or directories.</para>
  33         </listitem>
  34 <listitem>
  35           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296798" xreflabel=""/> Maintain a rough audit trail (file/directory changes with timestamps, but no user information).</para>
  36         </listitem>
  37 </itemizedlist>
  38       <para><anchor xml:id="dbdoclet.50438273_pgfId-1296879" xreflabel=""/>Changelogs record types are:</para>
  39       <informaltable frame="all">
  40         <tgroup cols="2">
  41           <colspec colname="c1" colwidth="50*"/>
  42           <colspec colname="c2" colwidth="50*"/>
  43           <thead>
  44             <row>
  45               <entry><para><emphasis role="bold"><anchor xml:id="dbdoclet.50438273_pgfId-1296808" xreflabel=""/>Value</emphasis></para></entry>
  46               <entry><para><emphasis role="bold"><anchor xml:id="dbdoclet.50438273_pgfId-1296810" xreflabel=""/>Description</emphasis></para></entry>
  47             </row>
  48           </thead>
  49           <tbody>
  50             <row>
  51               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296812" xreflabel=""/>MARK</para></entry>
  52               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296814" xreflabel=""/>Internal recordkeeping</para></entry>
  53             </row>
  54             <row>
  55               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296816" xreflabel=""/>CREAT</para></entry>
  56               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296818" xreflabel=""/>Regular file creation</para></entry>
  57             </row>
  58             <row>
  59               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296820" xreflabel=""/>MKDIR</para></entry>
  60               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296822" xreflabel=""/>Directory creation</para></entry>
  61             </row>
  62             <row>
  63               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296824" xreflabel=""/>HLINK</para></entry>
  64               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296826" xreflabel=""/>Hard link</para></entry>
  65             </row>
  66             <row>
  67               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296828" xreflabel=""/>SLINK</para></entry>
  68               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296830" xreflabel=""/>Soft link</para></entry>
  69             </row>
  70             <row>
  71               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296832" xreflabel=""/>MKNOD</para></entry>
  72               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296834" xreflabel=""/>Other file creation</para></entry>
  73             </row>
  74             <row>
  75               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296836" xreflabel=""/>UNLNK</para></entry>
  76               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296838" xreflabel=""/>Regular file removal</para></entry>
  77             </row>
  78             <row>
  79               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296840" xreflabel=""/>RMDIR</para></entry>
  80               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296842" xreflabel=""/>Directory removal</para></entry>
  81             </row>
  82             <row>
  83               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296844" xreflabel=""/>RNMFM</para></entry>
  84               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296846" xreflabel=""/>Rename, original</para></entry>
  85             </row>
  86             <row>
  87               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296848" xreflabel=""/>RNMTO</para></entry>
  88               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296850" xreflabel=""/>Rename, final</para></entry>
  89             </row>
  90             <row>
  91               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296860" xreflabel=""/>IOCTL</para></entry>
  92               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296862" xreflabel=""/>ioctl on file or directory</para></entry>
  93             </row>
  94             <row>
  95               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296864" xreflabel=""/>TRUNC</para></entry>
  96               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296866" xreflabel=""/>Regular file truncated</para></entry>
  97             </row>
  98             <row>
  99               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296868" xreflabel=""/>SATTR</para></entry>
 100               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296870" xreflabel=""/>Attribute change</para></entry>
 101             </row>
 102             <row>
 103               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296872" xreflabel=""/>XATTR</para></entry>
 104               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296874" xreflabel=""/>Extended attribute change</para></entry>
 105             </row>
 106             <row>
 107               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296876" xreflabel=""/>UNKNW</para></entry>
 108               <entry><para> <anchor xml:id="dbdoclet.50438273_pgfId-1296878" xreflabel=""/>Unknown operation</para></entry>
 109             </row>
 110           </tbody>
 111         </tgroup>
 112       </informaltable>
 113       <para><anchor xml:id="dbdoclet.50438273_pgfId-1296880" xreflabel=""/>FID-to-full-pathname and pathname-to-FID functions are also included to map target and parent FIDs into the file system namespace.</para>
 114       <section remap="h3">
 115         <title><anchor xml:id="dbdoclet.50438273_pgfId-1296889" xreflabel=""/>12.1.1 Working with Changelogs</title>
 116         <para><anchor xml:id="dbdoclet.50438273_pgfId-1296890" xreflabel=""/>Several commands are available to work with changelogs.</para>
 117         <section remap="h5">
 118           <title><anchor xml:id="dbdoclet.50438273_pgfId-1297845" xreflabel=""/>lctl changelog_register</title>
 119           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297846" xreflabel=""/>Because changelog records take up space on the MDT, the system administration must register changelog users. The registrants specify which records they are &quot;done with&quot;, and the system purges up to the greatest common record.</para>
 120           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297847" xreflabel=""/>To register a new changelog user, run:</para>
 121           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1297848" xreflabel=""/>lctl --device &lt;mdt_device&gt; changelog_register
 122 </screen>
 123           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297849" xreflabel=""/>Changelog entries are not purged beyond a registered user's set point (see lfs changelog_clear).</para>
 124         </section>
 125         <section remap="h5">
 126           <title><anchor xml:id="dbdoclet.50438273_pgfId-1297785" xreflabel=""/>lfs changelog</title>
 127           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297786" xreflabel=""/>To display the metadata changes on an MDT (the changelog records), run:</para>
 128           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1297825" xreflabel=""/>lfs changelog &lt;MDT name&gt; [startrec [endrec]]
 129 </screen>
 130           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297789" xreflabel=""/>It is optional whether to specify the start and end records.</para>
 131           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297836" xreflabel=""/>These are sample changelog records:</para>
 132           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1297790" xreflabel=""/>2 02MKDIR 4298396676 0x0 t=[0x200000405:0x15f9:0x0] p=[0x13:0x15e5a7a3:0x0]\
 133  pics
 134 <anchor xml:id="dbdoclet.50438273_pgfId-1297791" xreflabel=""/>3 01CREAT 4298402264 0x0 t=[0x200000405:0x15fa:0x0] p=[0x200000405:0x15f9:0\
 135 x0] chloe.jpg
 136 <anchor xml:id="dbdoclet.50438273_pgfId-1297792" xreflabel=""/>4 06UNLNK 4298404466 0x0 t=[0x200000405:0x15fa:0x0] p=[0x200000405:0x15f9:0\
 137 x0] chloe.jpg
 138 <anchor xml:id="dbdoclet.50438273_pgfId-1297793" xreflabel=""/>5 07RMDIR 4298405394 0x0 t=[0x200000405:0x15f9:0x0] p=[0x13:0x15e5a7a3:0x0]\
 139  pics
 140 </screen>
 141         </section>
 142         <section remap="h5">
 143           <title><anchor xml:id="dbdoclet.50438273_pgfId-1297809" xreflabel=""/>lfs changelog_clear</title>
 144           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297810" xreflabel=""/>To clear old changelog records for a specific user (records that the user no longer needs), run:</para>
 145           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1297811" xreflabel=""/>lfs changelog_clear &lt;MDT name&gt; &lt;user ID&gt; &lt;endrec&gt;
 146 </screen>
 147           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297812" xreflabel=""/>The changelog_clear command indicates that changelog records previous to &lt;endrec&gt; are no longer of interest to a particular user &lt;user ID&gt;, potentially allowing the MDT to free up disk space. An &lt;endrec&gt; value of 0 indicates the current last record. To run changelog_clear, the changelog user must be registered on the MDT node using lctl.</para>
 148           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297813" xreflabel=""/>When all changelog users are done with records &lt; X, the records are deleted.</para>
 149         </section>
 150         <section remap="h5">
 151           <title><anchor xml:id="dbdoclet.50438273_pgfId-1296930" xreflabel=""/>lctl changelog_deregister</title>
 152           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296931" xreflabel=""/>To deregister (unregister) a changelog user, run:</para>
 153           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1296932" xreflabel=""/>lctl --device &lt;mdt_device&gt; changelog_deregister &lt;user ID&gt;
 154 </screen>
 155           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297395" xreflabel=""/> Changelog_deregister cl1 effectively does a changelog_clear cl10 as it deregisters.</para>
 156         </section>
 157       </section>
 158       <section remap="h3">
 159         <title><anchor xml:id="dbdoclet.50438273_pgfId-1296934" xreflabel=""/>12.1.2 Changelog Examples</title>
 160         <para><anchor xml:id="dbdoclet.50438273_pgfId-1297340" xreflabel=""/>This section provides examples of different changelog commands.</para>
 161         <section remap="h5">
 162           <title><anchor xml:id="dbdoclet.50438273_pgfId-1297341" xreflabel=""/>Registering a Changelog User</title>
 163           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296935" xreflabel=""/>To register a new changelog user for a device (lustre-MDT0000):</para>
 164           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1296936" xreflabel=""/># lctl --device lustre-MDT0000 changelog_register
 165 <anchor xml:id="dbdoclet.50438273_pgfId-1296937" xreflabel=""/>lustre-MDT0000: Registered changelog userid &apos;cl1&apos;
 166 </screen>
 167         </section>
 168         <section remap="h5">
 169           <title><anchor xml:id="dbdoclet.50438273_pgfId-1296938" xreflabel=""/>Displaying Changelog Records</title>
 170           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296939" xreflabel=""/>To display changelog records on an MDT (lustre-MDT0000):</para>
 171           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1296940" xreflabel=""/>$ lfs changelog lustre-MDT0000
 172 <anchor xml:id="dbdoclet.50438273_pgfId-1296941" xreflabel=""/>1 00MARK  19:08:20.890432813 2010.03.24 0x0 t=[0x10001:0x0:0x0] p=[0:0x0:0x\
 173 0] mdd_obd-lustre-MDT0000-0
 174 <anchor xml:id="dbdoclet.50438273_pgfId-1296942" xreflabel=""/>2 02MKDIR 19:10:21.509659173 2010.03.24 0x0 t=[0x200000420:0x3:0x0] p=[0x61\
 175 b4:0xca2c7dde:0x0] mydir
 176 <anchor xml:id="dbdoclet.50438273_pgfId-1296943" xreflabel=""/>3 14SATTR 19:10:27.329356533 2010.03.24 0x0 t=[0x200000420:0x3:0x0]
 177 <anchor xml:id="dbdoclet.50438273_pgfId-1297627" xreflabel=""/>4 01CREAT 19:10:37.113847713 2010.03.24 0x0 t=[0x200000420:0x4:0x0] p=[0x20\
 178 0000420:0x3:0x0] hosts
 179 </screen>
 180           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297628" xreflabel=""/>Changelog records include this information:</para>
 181           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1297629" xreflabel=""/>rec#
 182 <anchor xml:id="dbdoclet.50438273_pgfId-1297630" xreflabel=""/>operation_type(numerical/text)
 183 <anchor xml:id="dbdoclet.50438273_pgfId-1297631" xreflabel=""/>timestamp
 184 <anchor xml:id="dbdoclet.50438273_pgfId-1297632" xreflabel=""/>datestamp
 185 <anchor xml:id="dbdoclet.50438273_pgfId-1297633" xreflabel=""/>flags
 186 <anchor xml:id="dbdoclet.50438273_pgfId-1297634" xreflabel=""/>t=target_FID
 187 <anchor xml:id="dbdoclet.50438273_pgfId-1297635" xreflabel=""/>p=parent_FID
 188 <anchor xml:id="dbdoclet.50438273_pgfId-1297636" xreflabel=""/>target_name
 189 </screen>
 190           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297637" xreflabel=""/>Displayed in this format:</para>
 191           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1297638" xreflabel=""/>rec# operation_type(numerical/text) timestamp datestamp flags t=target_FID \
 192 p=parent_FID target_name
 193 </screen>
 194           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297639" xreflabel=""/>For example:</para>
 195           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1297640" xreflabel=""/>4 01CREAT 19:10:37.113847713 2010.03.24 0x0 t=[0x200000420:0x4:0x0] p=[0x20\
 196 0000420:0x3:0x0] hosts
 197 </screen>
 198         </section>
 199         <section remap="h5">
 200           <title><anchor xml:id="dbdoclet.50438273_pgfId-1297641" xreflabel=""/>Clearing Changelog Records</title>
 201           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297348" xreflabel=""/>To notify a device that a specific user (cl1) no longer needs records (up to and including 3):</para>
 202           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1296947" xreflabel=""/>$ lfs changelog_clear  lustre-MDT0000 cl1 3
 203 </screen>
 204           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296958" xreflabel=""/>To confirm that the changelog_clear operation was successful, run lfs changelog; only records after id-3 are listed:</para>
 205           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1297062" xreflabel=""/>$ lfs changelog lustre-MDT0000
 206 <anchor xml:id="dbdoclet.50438273_pgfId-1297063" xreflabel=""/>4 01CREAT 19:10:37.113847713 2010.03.24 0x0 t=[0x200000420:0x4:0x0] p=[0x20\
 207 0000420:0x3:0x0] hosts
 208 </screen>
 209         </section>
 210         <section remap="h5">
 211           <title><anchor xml:id="dbdoclet.50438273_pgfId-1297051" xreflabel=""/>Deregistering a Changelog User</title>
 212           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297030" xreflabel=""/>To deregister a changelog user (cl1) for a specific device (lustre-MDT0000):</para>
 213           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1296959" xreflabel=""/># lctl --device lustre-MDT0000 changelog_deregister cl1
 214 <anchor xml:id="dbdoclet.50438273_pgfId-1296960" xreflabel=""/>lustre-MDT0000: Deregistered changelog user &apos;cl1&apos;
 215 </screen>
 216           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296961" xreflabel=""/>The deregistration operation clears all changelog records for the specified user (cli).</para>
 217           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1296962" xreflabel=""/>$ lfs changelog lustre-MDT0000
 218 <anchor xml:id="dbdoclet.50438273_pgfId-1296963" xreflabel=""/>5 00MARK  19:13:40.858292517 2010.03.24 0x0 t=[0x40001:0x0:0x0] p=[0:0x0:0x\
 219 0] mdd_obd-lustre-MDT0000-0
 220 </screen>
 221           <informaltable frame="none">
 222             <tgroup cols="1">
 223               <colspec colname="c1" colwidth="100*"/>
 224               <tbody>
 225                 <row>
 226                   <entry><para><emphasis role="bold">Note -</emphasis><anchor xml:id="dbdoclet.50438273_pgfId-1296965" xreflabel=""/>MARK records typically indicate changelog recording status changes.</para></entry>
 227                 </row>
 228               </tbody>
 229             </tgroup>
 230           </informaltable>
 231         </section>
 232         <section remap="h5">
 233           <title><anchor xml:id="dbdoclet.50438273_pgfId-1296966" xreflabel=""/>Displaying the Changelog Index and Registered Users</title>
 234           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296971" xreflabel=""/>To display the current, maximum changelog index and registered changelog users for a specific device (lustre-MDT0000):</para>
 235           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1296972" xreflabel=""/># lctl get_param  mdd.lustre-MDT0000.changelog_users
 236 <anchor xml:id="dbdoclet.50438273_pgfId-1296973" xreflabel=""/>mdd.lustre-MDT0000.changelog_users=current index: 8
 237 <anchor xml:id="dbdoclet.50438273_pgfId-1296974" xreflabel=""/>ID    index
 238 <anchor xml:id="dbdoclet.50438273_pgfId-1296975" xreflabel=""/>cl2   8
 239 </screen>
 240         </section>
 241         <section remap="h5">
 242           <title><anchor xml:id="dbdoclet.50438273_pgfId-1296976" xreflabel=""/>Displaying the Changelog Mask</title>
 243           <para><anchor xml:id="dbdoclet.50438273_pgfId-1296977" xreflabel=""/>To show the current changelog mask on a specific device (lustre-MDT0000):</para>
 244           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1296978" xreflabel=""/># lctl get_param  mdd.lustre-MDT0000.changelog_mask
 245 <anchor xml:id="dbdoclet.50438273_pgfId-1296979" xreflabel=""/>mdd.lustre-MDT0000.changelog_mask=
 246 <anchor xml:id="dbdoclet.50438273_pgfId-1296980" xreflabel=""/>MARK CREAT MKDIR HLINK SLINK MKNOD UNLNK RMDIR RNMFM RNMTO OPEN CLOSE IOCTL\
 247  TRUNC SATTR XATTR HSM
 248 </screen>
 249         </section>
 250         <section remap="h5">
 251           <title><anchor xml:id="dbdoclet.50438273_pgfId-1297457" xreflabel=""/>Setting the Changelog Mask</title>
 252           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297458" xreflabel=""/>To set the current changelog mask on a specific device (lustre-MDT0000):</para>
 253           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1297459" xreflabel=""/># lctl set_param mdd.lustre-MDT0000.changelog_mask=HLINK
 254 <anchor xml:id="dbdoclet.50438273_pgfId-1297460" xreflabel=""/>mdd.lustre-MDT0000.changelog_mask=HLINK
 255 <anchor xml:id="dbdoclet.50438273_pgfId-1297288" xreflabel=""/>$ lfs changelog_clear lustre-MDT0000 cl1 0
 256 <anchor xml:id="dbdoclet.50438273_pgfId-1297293" xreflabel=""/>$ mkdir /mnt/lustre/mydir/foo
 257 <anchor xml:id="dbdoclet.50438273_pgfId-1297314" xreflabel=""/>$ cp /etc/hosts /mnt/lustre/mydir/foo/file
 258 <anchor xml:id="dbdoclet.50438273_pgfId-1297316" xreflabel=""/>$ ln /mnt/lustre/mydir/foo/file /mnt/lustre/mydir/myhardlink
 259 </screen>
 260           <para><anchor xml:id="dbdoclet.50438273_pgfId-1297318" xreflabel=""/> Only item types that are in the mask show up in the changelog.</para>
 261           <screen><anchor xml:id="dbdoclet.50438273_pgfId-1297441" xreflabel=""/>$ lfs changelog lustre-MDT0000
 262 <anchor xml:id="dbdoclet.50438273_pgfId-1297321" xreflabel=""/>9 03HLINK 19:19:35.171867477 2010.03.24 0x0 t=[0x200000420:0x6:0x0] p=[0x20\
 263 0000420:0x3:0x0] myhardlink
 264 </screen>
 265         </section>
 266       </section>
 267     </section>
 268     <section xml:id="dbdoclet.50438273_81684">
 269       <title>12.2 Lustre <anchor xml:id="dbdoclet.50438273_marker-1297386" xreflabel=""/>Monitoring Tool</title>
 270       <para><anchor xml:id="dbdoclet.50438273_pgfId-1298196" xreflabel=""/>The Lustre Monitoring Tool (LMT) is a Python-based, distributed system developed and maintained by Lawrence Livermore National Lab (LLNL)). It provides a &apos;&apos;top&apos;&apos; like display of activity on server-side nodes (MDS, OSS and portals routers) on one or more Lustre file systems. It does not provide support for monitoring clients. For more information on LMT, including the setup procedure, see:</para>
 271       <para><anchor xml:id="dbdoclet.50438273_pgfId-1296057" xreflabel=""/><link xl:href="http://code.google.com/p/lmt/">http://code.google.com/p/lmt/</link></para>
 272       <para><anchor xml:id="dbdoclet.50438273_pgfId-1296102" xreflabel=""/>LMT questions can be directed to:</para>
 273       <para><anchor xml:id="dbdoclet.50438273_pgfId-1296120" xreflabel=""/><link xl:href="mailto:lmt-discuss@googlegroups.com">lmt-discuss@googlegroups.com</link></para>
 274     </section>
 275     <section xml:id="dbdoclet.50438273_80593">
 276       <title>12.3 Collect<anchor xml:id="dbdoclet.50438273_marker-1297391" xreflabel=""/>L</title>
 277       <para><anchor xml:id="dbdoclet.50438273_pgfId-1293325" xreflabel=""/>CollectL is another tool that can be used to monitor Lustre. You can run CollectL on a Lustre system that has any combination of MDSs, OSTs and clients. The collected data can be written to a file for continuous logging and played back at a later time. It can also be converted to a format suitable for plotting.</para>
 278       <para><anchor xml:id="dbdoclet.50438273_pgfId-1293329" xreflabel=""/>For more information about CollectL, see:</para>
 279       <para><anchor xml:id="dbdoclet.50438273_pgfId-1294650" xreflabel=""/><link xl:href="http://collectl.sourceforge.net">http://collectl.sourceforge.net</link></para>
 280       <para><anchor xml:id="dbdoclet.50438273_pgfId-1294692" xreflabel=""/>Lustre-specific documentation is also available. See:</para>
 281       <para><anchor xml:id="dbdoclet.50438273_pgfId-1294696" xreflabel=""/><link xl:href="http://collectl.sourceforge.net/Tutorial-Lustre.html">http://collectl.sourceforge.net/Tutorial-Lustre.html</link></para>
 282     </section>
 283     <section xml:id="dbdoclet.50438273_44185">
 284       <title>12.4 Other Monitoring Options</title>
 285       <para><anchor xml:id="dbdoclet.50438273_pgfId-1295490" xreflabel=""/>A variety of standard tools are available publically.</para>
 286       <para><anchor xml:id="dbdoclet.50438273_pgfId-1298235" xreflabel=""/>Another option is to script a simple monitoring solution that looks at various reports from ipconfig, as well as the procfs files generated by Lustre.</para>
 287       <para><anchor xml:id="dbdoclet.50438273_pgfId-1296750" xreflabel=""/><anchor xml:id="dbdoclet.50438273_67514" xreflabel=""/> </para>
 288     </section>
 289 </chapter>