1 <?xml version='1.0' encoding='UTF-8'?><chapter xmlns="http://docbook.org/ns/docbook" xmlns:xl="http://www.w3.org/1999/xlink" version="5.0" xml:lang="en-US" xml:id="lustremaintenance">
2 <title xml:id="lustremaintenance.title">Lustre Maintenance</title>
3 <para>Once you have the Lustre file system up and running, you can use the procedures in this section to perform these basic Lustre maintenance tasks:</para>
6 <para><xref linkend="dbdoclet.50438199_42877"/></para>
9 <para><xref linkend="dbdoclet.50438199_15240"/></para>
12 <para><xref linkend="dbdoclet.50438199_26070"/></para>
15 <para><xref linkend="dbdoclet.50438199_54623"/></para>
18 <para><xref linkend="dbdoclet.changingservernid"/></para>
21 <para><xref linkend="dbdoclet.adding_new_mdt"/></para>
24 <para><xref linkend="dbdoclet.adding_new_ost"/></para>
27 <para><xref linkend="dbdoclet.deactivating_mdt_ost"/></para>
30 <para><xref linkend="dbdoclet.rmremotedir"/></para>
33 <para><xref linkend="dbdoclet.inactivemdt"/></para>
36 <para><xref xmlns:xlink="http://www.w3.org/1999/xlink" linkend="section_remove_ost"/></para>
39 <para><xref xmlns:xlink="http://www.w3.org/1999/xlink" linkend="section_ydg_pgt_tl"/></para>
42 <para><xref xmlns:xlink="http://www.w3.org/1999/xlink" linkend="section_restore_ost"/></para>
45 <para><xref xmlns:xlink="http://www.w3.org/1999/xlink" linkend="section_ucf_qgt_tl"/></para>
48 <para><xref linkend="dbdoclet.50438199_77819"/></para>
51 <para><xref linkend="dbdoclet.50438199_12607"/></para>
54 <para><xref linkend="dbdoclet.50438199_62333"/></para>
57 <para><xref linkend="dbdoclet.50438199_62545"/></para>
60 <section xml:id="dbdoclet.50438199_42877">
62 <indexterm><primary>maintenance</primary></indexterm>
63 <indexterm><primary>maintenance</primary><secondary>inactive OSTs</secondary></indexterm>
64 Working with Inactive OSTs</title>
65 <para>To mount a client or an MDT with one or more inactive OSTs, run commands similar to this:</para>
66 <screen>client# mount -o exclude=testfs-OST0000 -t lustre \
67 uml1:/testfs /mnt/testfs
68 client# lctl get_param lov.testfs-clilov-*.target_obd</screen>
69 <para>To activate an inactive OST on a live client or MDT, use the
70 <literal>lctl activate</literal> command on the OSC device. For example:</para>
71 <screen>lctl --device 7 activate</screen>
73 <para>A colon-separated list can also be specified. For example,
74 <literal>exclude=testfs-OST0000:testfs-OST0001</literal>.</para>
77 <section xml:id="dbdoclet.50438199_15240">
78 <title><indexterm><primary>maintenance</primary><secondary>finding nodes</secondary></indexterm>
79 Finding Nodes in the Lustre File System</title>
80 <para>There may be situations in which you need to find all nodes in
81 your Lustre file system or get the names of all OSTs.</para>
82 <para>To get a list of all Lustre nodes, run this command on the MGS:</para>
83 <screen># lctl get_param mgs.MGS.live.*</screen>
85 <para>This command must be run on the MGS.</para>
87 <para>In this example, file system <literal>testfs</literal> has three
88 nodes, <literal>testfs-MDT0000</literal>,
89 <literal>testfs-OST0000</literal>, and
90 <literal>testfs-OST0001</literal>.</para>
91 <screen>mgs:/root# lctl get_param mgs.MGS.live.*
96 testfs-OST0001 </screen>
97 <para>To get the names of all OSTs, run this command on the MDS:</para>
98 <screen>mds:/root# lctl get_param lov.*-mdtlov.target_obd </screen>
100 <para>This command must be run on the MDS.</para>
102 <para>In this example, there are two OSTs, testfs-OST0000 and
103 testfs-OST0001, which are both active.</para>
104 <screen>mgs:/root# lctl get_param lov.testfs-mdtlov.target_obd
105 0: testfs-OST0000_UUID ACTIVE
106 1: testfs-OST0001_UUID ACTIVE </screen>
108 <section xml:id="dbdoclet.50438199_26070">
109 <title><indexterm><primary>maintenance</primary><secondary>mounting a server</secondary></indexterm>
110 Mounting a Server Without Lustre Service</title>
111 <para>If you are using a combined MGS/MDT, but you only want to start the MGS and not the MDT, run this command:</para>
112 <screen>mount -t lustre <replaceable>/dev/mdt_partition</replaceable> -o nosvc <replaceable>/mount_point</replaceable></screen>
113 <para>The <literal><replaceable>mdt_partition</replaceable></literal> variable is the combined MGS/MDT block device.</para>
114 <para>In this example, the combined MGS/MDT is <literal>testfs-MDT0000</literal> and the mount point is <literal>/mnt/test/mdt</literal>.</para>
115 <screen>$ mount -t lustre -L testfs-MDT0000 -o nosvc /mnt/test/mdt</screen>
117 <section xml:id="dbdoclet.50438199_54623">
118 <title><indexterm><primary>maintenance</primary><secondary>regenerating config logs</secondary></indexterm>
119 Regenerating Lustre Configuration Logs</title>
120 <para>If the Lustre file system configuration logs are in a state where the file system cannot
121 be started, use the <literal>writeconf</literal> command to erase them. After the
122 <literal>writeconf</literal> command is run and the servers restart, the configuration logs
123 are re-generated and stored on the MGS (as in a new file system).</para>
124 <para>You should only use the <literal>writeconf</literal> command if:</para>
127 <para>The configuration logs are in a state where the file system cannot start</para>
130 <para>A server NID is being changed</para>
133 <para>The <literal>writeconf</literal> command is destructive to some configuration items (i.e., OST pools information and items set via <literal>conf_param</literal>), and should be used with caution. To avoid problems:</para>
136 <para>Shut down the file system before running the <literal>writeconf</literal> command</para>
139 <para>Run the <literal>writeconf</literal> command on all servers (MDT first, then OSTs)</para>
142 <para>Start the file system in this order:</para>
145 <para>MGS (or the combined MGS/MDT)</para>
154 <para>Lustre clients</para>
160 <para>The OST pools feature enables a group of OSTs to be named for file striping purposes. If you use OST pools, be aware that running the <literal>writeconf</literal> command erases <emphasis role="bold">all</emphasis> pools information (as well as any other parameters set via <literal>lctl conf_param</literal>). We recommend that the pools definitions (and <literal>conf_param</literal> settings) be executed via a script, so they can be reproduced easily after a <literal>writeconf</literal> is performed.</para>
162 <para>To regenerate Lustre file system configuration logs:</para>
165 <para>Shut down the file system in this order.</para>
168 <para>Unmount the clients.</para>
171 <para>Unmount the MDT.</para>
174 <para>Unmount all OSTs.</para>
179 <para>Make sure the the MDT and OST devices are available.</para>
182 <para>Run the <literal>writeconf</literal> command on all servers.</para>
183 <para>Run writeconf on the MDT first, and then the OSTs.</para>
186 <para>On the MDT, run:</para>
187 <screen>mdt# tunefs.lustre --writeconf <replaceable>/dev/mdt_device</replaceable></screen>
193 <screen>ost# tunefs.lustre --writeconf <replaceable>/dev/ost_device</replaceable></screen>
199 <para>Restart the file system in this order.</para>
202 <para>Mount the MGS (or the combined MGS/MDT).</para>
205 <para>Mount the MDT.</para>
208 <para>Mount the OSTs.</para>
211 <para>Mount the clients.</para>
216 <para>After the <literal>writeconf</literal> command is run, the configuration logs are re-generated as servers restart.</para>
218 <section xml:id="dbdoclet.changingservernid">
219 <title><indexterm><primary>maintenance</primary><secondary>changing a NID</secondary></indexterm>
220 Changing a Server NID</title>
221 <para>In Lustre software release 2.3 or earlier, the <literal>tunefs.lustre
222 --writeconf</literal> command is used to rewrite all of the configuration files.</para>
223 <para condition="l24">If you need to change the NID on the MDT or OST, a new
224 <literal>replace_nids</literal> command was added in Lustre software release 2.4 to simplify
225 this process. The <literal>replace_nids</literal> command differs from <literal>tunefs.lustre
226 --writeconf</literal> in that it does not erase the entire configuration log, precluding the
227 need the need to execute the <literal>writeconf</literal> command on all servers and
228 re-specify all permanent parameter settings. However, the <literal>writeconf</literal> command
229 can still be used if desired.</para>
230 <para>Change a server NID in these situations:</para>
233 <para>New server hardware is added to the file system, and the MDS or an OSS is being moved to the new machine.</para>
236 <para>New network card is installed in the server.</para>
239 <para>You want to reassign IP addresses.</para>
242 <para>To change a server NID:</para>
245 <para>Update the LNet configuration in the <literal>/etc/modprobe.conf</literal> file so the list of server NIDs is correct. Use <literal>lctl list_nids</literal> to view the list of server NIDS.</para>
246 <para>The <literal>lctl list_nids</literal> command indicates which network(s) are
247 configured to work with the Lustre file system.</para>
250 <para>Shut down the file system in this order:</para>
253 <para>Unmount the clients.</para>
256 <para>Unmount the MDT.</para>
259 <para>Unmount all OSTs.</para>
264 <para>If the MGS and MDS share a partition, start the MGS only:</para>
265 <screen>mount -t lustre <replaceable>MDT partition</replaceable> -o nosvc <replaceable>mount_point</replaceable></screen>
268 <para>Run the <literal>replace_nids</literal> command on the MGS:</para>
269 <screen>lctl replace_nids <replaceable>devicename</replaceable> <replaceable>nid1</replaceable>[,nid2,nid3 ...]</screen>
270 <para>where <replaceable>devicename</replaceable> is the Lustre target name, e.g.
271 <literal>testfs-OST0013</literal></para>
274 <para>If the MGS and MDS share a partition, stop the MGS:</para>
275 <screen>umount <replaceable>mount_point</replaceable></screen>
278 <note><para>The <literal>replace_nids</literal> command also cleans
279 all old, invalidated records out of the configuration log, while
280 preserving all other current settings.</para></note>
281 <note><para>The previous configuration log is backed up on the MGS
282 disk with the suffix <literal>'.bak'</literal>.</para></note>
284 <section xml:id="dbdoclet.adding_new_mdt" condition='l24'>
286 <primary>maintenance</primary>
287 <secondary>adding an MDT</secondary>
288 </indexterm>Adding a New MDT to a Lustre File System</title>
289 <para>Additional MDTs can be added using the DNE feature to serve one
290 or more remote sub-directories within a filesystem, in order to
291 increase the total number of files that can be created in the
292 filesystem, to increase aggregate metadata performance, or to isolate
293 user or application workloads from other users of the filesystem. It
294 is possible to have multiple remote sub-directories reference the
295 same MDT. However, the root directory will always be located on
296 MDT0. To add a new MDT into the file system:</para>
299 <para>Discover the maximum MDT index. Each MDT must have unique index.</para>
301 client$ lctl dl | grep mdc
302 36 UP mdc testfs-MDT0000-mdc-ffff88004edf3c00 4c8be054-144f-9359-b063-8477566eb84e 5
303 37 UP mdc testfs-MDT0001-mdc-ffff88004edf3c00 4c8be054-144f-9359-b063-8477566eb84e 5
304 38 UP mdc testfs-MDT0002-mdc-ffff88004edf3c00 4c8be054-144f-9359-b063-8477566eb84e 5
305 39 UP mdc testfs-MDT0003-mdc-ffff88004edf3c00 4c8be054-144f-9359-b063-8477566eb84e 5
309 <para>Add the new block device as a new MDT at the next available
310 index. In this example, the next available index is 4.</para>
312 mds# mkfs.lustre --reformat --fsname=<replaceable>testfs</replaceable> --mdt --mgsnode=<replaceable>mgsnode</replaceable> --index 4 <replaceable>/dev/mdt4_device</replaceable>
316 <para>Mount the MDTs.</para>
318 mds# mount –t lustre <replaceable>/dev/mdt4_blockdevice</replaceable> /mnt/mdt4
322 <para>In order to start creating new files and directories on the
323 new MDT(s) they need to be attached into the namespace at one or
324 more subdirectories using the <literal>lfs mkdir</literal> command.
325 All files and directories below those created with
326 <literal>lfs mkdir</literal> will also be created on the same MDT
327 unless otherwise specified.
330 client# lfs mkdir -i 3 /mnt/testfs/new_dir_on_mdt3
331 client# lfs mkdir -i 4 /mnt/testfs/new_dir_on_mdt4
332 client# lfs mkdir -c 4 /mnt/testfs/new_directory_striped_across_4_mdts
337 <section xml:id="dbdoclet.adding_new_ost">
338 <title><indexterm><primary>maintenance</primary><secondary>adding a OST</secondary></indexterm>
339 Adding a New OST to a Lustre File System</title>
340 <para>A new OST can be added to existing Lustre file system on either
341 an existing OSS node or on a new OSS node. In order to keep client IO
342 load balanced across OSS nodes for maximum aggregate performance, it is
343 not recommended to configure different numbers of OSTs to each OSS node.
347 <para> Add a new OST by using <literal>mkfs.lustre</literal> as when
348 the filesystem was first formatted, see
349 <xref linkend="dbdoclet.format_ost" /> for details. Each new OST
350 must have a unique index number, use <literal>lctl dl</literal> to
351 see a list of all OSTs. For example, to add a new OST at index 12
352 to the <literal>testfs</literal> filesystem run following commands
353 should be run on the OSS:</para>
354 <screen>oss# mkfs.lustre --fsname=testfs --mgsnode=mds16@tcp0 --ost --index=12 /dev/sda
355 oss# mkdir -p /mnt/testfs/ost12
356 oss# mount -t lustre /dev/sda /mnt/testfs/ost12</screen>
359 <para>Balance OST space usage (possibly).</para>
360 <para>The file system can be quite unbalanced when new empty OSTs
361 are added to a relatively full filesystem. New file creations are
362 automatically balanced to favour the new OSTs. If this is a scratch
363 file system or files are pruned at regular intervals, then no further
364 work may be needed to balance the OST space usage as new files being
365 created will preferentially be placed on the less full OST(s). As old
366 files are deleted, they will release space on the old OST(s).</para>
367 <para>Files existing prior to the expansion can optionally be
368 rebalanced using the <literal>lfs_migrate</literal> utility.
369 This redistributes file data over the entire set of OSTs.</para>
370 <para>For example, to rebalance all files within the directory
371 <literal>/mnt/lustre/dir</literal>, enter:</para>
372 <screen>client# lfs_migrate /mnt/lustre/file</screen>
373 <para>To migrate files within the <literal>/test</literal> file
374 system on <literal>OST0004</literal> that are larger than 4GB in
375 size to other OSTs, enter:</para>
376 <screen>client# lfs find /test --ost test-OST0004 -size +4G | lfs_migrate -y</screen>
377 <para>See <xref linkend="dbdoclet.lfs_migrate"/> for details.</para>
381 <section xml:id="dbdoclet.deactivating_mdt_ost">
382 <title><indexterm><primary>maintenance</primary><secondary>restoring an OST</secondary></indexterm>
383 <indexterm><primary>maintenance</primary><secondary>removing an OST</secondary></indexterm>
384 Removing and Restoring MDTs and OSTs</title>
385 <para>OSTs and DNE MDTs can be removed from and restored to a Lustre
386 filesystem. Deactivating an OST means that it is temporarily or
387 permanently marked unavailable. Deactivating an OST on the MDS means
388 it will not try to allocate new objects there or perform OST recovery,
389 while deactivating an OST the client means it will not wait for OST
390 recovery if it cannot contact the OST and will instead return an IO
391 error to the application immediately if files on the OST are accessed.
392 An OST may be permanently deactivated from the file system,
393 depending on the situation and commands used.</para>
394 <note><para>A permanently deactivated MDT or OST still appears in the
395 filesystem configuration until the configuration is regenerated with
396 <literal>writeconf</literal> or it is replaced with a new MDT or OST
397 at the same index and permanently reactivated. A deactivated OST
398 will not be listed by <literal>lfs df</literal>.
400 <para>You may want to temporarily deactivate an OST on the MDS to
401 prevent new files from being written to it in several situations:</para>
404 <para>A hard drive has failed and a RAID resync/rebuild is underway,
405 though the OST can also be marked <emphasis>degraded</emphasis> by
406 the RAID system to avoid allocating new files on the slow OST which
407 can reduce performance, see <xref linkend='dbdoclet.degraded_ost' />
412 <para>OST is nearing its space capacity, though the MDS will already
413 try to avoid allocating new files on overly-full OSTs if possible,
414 see <xref linkend='dbdoclet.balancing_free_space' /> for details.
418 <para>MDT/OST storage or MDS/OSS node has failed, and will not
419 be available for some time (or forever), but there is still a
420 desire to continue using the filesystem before it is repaired.</para>
423 <section condition="l24" xml:id="dbdoclet.rmremotedir">
424 <title><indexterm><primary>maintenance</primary><secondary>removing an MDT</secondary></indexterm>Removing an MDT from the File System</title>
425 <para>If the MDT is permanently inaccessible,
426 <literal>lfs rm_entry {directory}</literal> can be used to delete the
427 directory entry for the unavailable MDT. Using <literal>rmdir</literal>
428 would otherwise report an IO error due to the remote MDT being inactive.
429 Please note that if the MDT <emphasis>is</emphasis> available, standard
430 <literal>rm -r</literal> should be used to delete the remote directory.
431 After the remote directory has been removed, the administrator should
432 mark the MDT as permanently inactive with:</para>
433 <screen>lctl conf_param {MDT name}.mdc.active=0</screen>
434 <para>A user can identify which MDT holds a remote sub-directory using
435 the <literal>lfs</literal> utility. For example:</para>
436 <screen>client$ lfs getstripe --mdt-index /mnt/lustre/remote_dir1
438 client$ mkdir /mnt/lustre/local_dir0
439 client$ lfs getstripe --mdt-index /mnt/lustre/local_dir0
442 <para>The <literal>lfs getstripe --mdt-index</literal> command
443 returns the index of the MDT that is serving the given directory.</para>
445 <section xml:id="dbdoclet.inactivemdt" condition='l24'>
447 <indexterm><primary>maintenance</primary></indexterm>
448 <indexterm><primary>maintenance</primary><secondary>inactive MDTs</secondary></indexterm>Working with Inactive MDTs</title>
449 <para>Files located on or below an inactive MDT are inaccessible until
450 the MDT is activated again. Clients accessing an inactive MDT will receive
453 <section remap="h3" xml:id="section_remove_ost">
455 <primary>maintenance</primary>
456 <secondary>removing an OST</secondary>
457 </indexterm>Removing an OST from the File System</title>
458 <para>When deactivating an OST, note that the client and MDS each have
459 an OSC device that handles communication with the corresponding OST.
460 To remove an OST from the file system:</para>
463 <para>If the OST is functional, and there are files located on
464 the OST that need to be migrated off of the OST, the file creation
465 for that OST should be temporarily deactivated on the MDS (each MDS
466 if running with multiple MDS nodes in DNE mode).
470 <para condition="l29">With Lustre 2.9 and later, the MDS should be
471 set to only disable file creation on that OST by setting
472 <literal>max_create_count</literal> to zero:
473 <screen>mds# lctl set_param osp.<replaceable>osc_name</replaceable>.max_create_count=0</screen>
474 This ensures that files deleted or migrated off of the OST
475 will have their corresponding OST objects destroyed, and the space
476 will be freed. For example, to disable <literal>OST0000</literal>
477 in the filesystem <literal>testfs</literal>, run:
478 <screen>mds# lctl set_param osp.testfs-OST0000-osc-MDT*.max_create_count=0</screen>
479 on each MDS in the <literal>testfs</literal> filesystem.</para>
482 <para>With older versions of Lustre, to deactivate the OSC on the
484 <screen>mds# lctl set_param osp.<replaceable>osc_name</replaceable>.active=0</screen>
485 This will prevent the MDS from attempting any communication with
486 that OST, including destroying objects located thereon. This is
487 fine if the OST will be removed permanently, if the OST is not
488 stable in operation, or if it is in a read-only state. Otherwise,
489 the free space and objects on the OST will not decrease when
490 files are deleted, and object destruction will be deferred until
491 the MDS reconnects to the OST.</para>
492 <para>For example, to deactivate <literal>OST0000</literal> in
493 the filesystem <literal>testfs</literal>, run:
494 <screen>mds# lctl set_param osp.testfs-OST0000-osc-MDT*.active=0</screen>
495 Deactivating the OST on the <emphasis>MDS</emphasis> does not
496 prevent use of existing objects for read/write by a client.</para>
498 <para>If migrating files from a working OST, do not deactivate
499 the OST on clients. This causes IO errors when accessing files
500 located there, and migrating files on the OST would fail.</para>
503 <para>Do not use <literal>lctl conf_param</literal> to
504 deactivate the OST if it is still working, as this immediately
505 and permanently deactivates it in the file system configuration
506 on both the MDS and all clients.</para>
512 <para>Discover all files that have objects residing on the
513 deactivated OST. Depending on whether the deactivated OST is
514 available or not, the data from that OST may be migrated to
515 other OSTs, or may need to be restored from backup.</para>
518 <para>If the OST is still online and available, find all
519 files with objects on the deactivated OST, and copy them
520 to other OSTs in the file system to: </para>
521 <screen>client# lfs find --ost <replaceable>ost_name</replaceable> <replaceable>/mount/point</replaceable> | lfs_migrate -y</screen>
524 <para>If the OST is no longer available, delete the files
525 on that OST and restore them from backup:
526 <screen>client# lfs find --ost <replaceable>ost_uuid</replaceable> -print0 <replaceable>/mount/point</replaceable> |
527 tee /tmp/files_to_restore | xargs -0 -n 1 unlink</screen>
528 The list of files that need to be restored from backup is
529 stored in <literal>/tmp/files_to_restore</literal>. Restoring
530 these files is beyond the scope of this document.</para>
535 <para>Deactivate the OST.</para>
539 If there is expected to be a replacement OST in some short
540 time (a few days), the OST can temporarily be deactivated on
542 <screen>client# lctl set_param osc.<replaceable>fsname</replaceable>-OST<replaceable>number</replaceable>-*.active=0</screen>
543 <note><para>This setting is only temporary and will be reset
544 if the clients are remounted or rebooted. It needs to be run
545 on all clients.</para>
550 <para>If there is not expected to be a replacement for this OST in
551 the near future, permanently deactivate it on all clients and
552 the MDS by running the following command on the MGS:
553 <screen>mgs# lctl conf_param <replaceable>ost_name</replaceable>.osc.active=0</screen></para>
554 <note><para>A deactivated OST still appears in the file system
555 configuration, though a replacement OST can be created using the
556 <literal>mkfs.lustre --replace</literal> option, see
557 <xref linkend="section_restore_ost"/>.
564 <section remap="h3" xml:id="section_ydg_pgt_tl">
566 <primary>maintenance</primary>
567 <secondary>backing up OST config</secondary>
570 <primary>backup</primary>
571 <secondary>OST config</secondary>
572 </indexterm> Backing Up OST Configuration Files</title>
573 <para>If the OST device is still accessible, then the Lustre
574 configuration files on the OST should be backed up and saved for
575 future use in order to avoid difficulties when a replacement OST is
576 returned to service. These files rarely change, so they can and
577 should be backed up while the OST is functional and accessible. If
578 the deactivated OST is still available to mount (i.e. has not
579 permanently failed or is unmountable due to severe corruption), an
580 effort should be made to preserve these files. </para>
583 <para>Mount the OST file system.
584 <screen>oss# mkdir -p /mnt/ost
585 oss# mount -t ldiskfs <replaceable>/dev/ost_device</replaceable> /mnt/ost</screen>
589 <para>Back up the OST configuration files.
590 <screen>oss# tar cvf <replaceable>ost_name</replaceable>.tar -C /mnt/ost last_rcvd \
591 CONFIGS/ O/0/LAST_ID</screen>
595 <para> Unmount the OST file system. <screen>oss# umount /mnt/ost</screen>
600 <section xml:id="section_restore_ost">
602 <primary>maintenance</primary>
603 <secondary>restoring OST config</secondary>
606 <primary>backup</primary>
607 <secondary>restoring OST config</secondary>
608 </indexterm> Restoring OST Configuration Files</title>
609 <para>If the original OST is still available, it is best to follow the
610 OST backup and restore procedure given in either
611 <xref linkend="dbdoclet.backup_device"/>, or
612 <xref linkend="backup_fs_level"/> and
613 <xref linkend="backup_fs_level.restore"/>.</para>
614 <para>To replace an OST that was removed from service due to corruption
615 or hardware failure, the replacement OST needs to be formatted using
616 <literal>mkfs.lustre</literal>, and the Lustre file system configuration
617 should be restored, if available. Any objects stored on the OST will
618 be permanently lost, and files using the OST should be deleted and/or
619 restored from backup.</para>
620 <para condition="l25">With Lustre 2.5 and later, it is possible to
621 replace an OST to the same index without restoring the configuration
622 files, using the <literal>--replace</literal> option at format time.
623 <screen>oss# mkfs.lustre --ost --reformat --replace --index=<replaceable>old_ost_index</replaceable> \
624 <replaceable>other_options</replaceable> <replaceable>/dev/new_ost_dev</replaceable></screen>
625 The MDS and OSS will negotiate the <literal>LAST_ID</literal> value
626 for the replacement OST.
628 <para>If the OST configuration files were not backed up, due to the
629 OST file system being completely inaccessible, it is still possible to
630 replace the failed OST with a new one at the same OST index. </para>
633 <para>For older versions, format the OST file system without the
634 <literal>--replace</literal> option and restore the saved
636 <screen>oss# mkfs.lustre --ost --reformat --index=<replaceable>old_ost_index</replaceable> \
637 <replaceable>other_options</replaceable> <replaceable>/dev/new_ost_dev</replaceable></screen>
641 <para> Mount the OST file system.
642 <screen>oss# mkdir /mnt/ost
643 oss# mount -t ldiskfs <replaceable>/dev/new_ost_dev</replaceable> <replaceable>/mnt/ost</replaceable></screen>
647 <para>Restore the OST configuration files, if available.
648 <screen>oss# tar xvf <replaceable>ost_name</replaceable>.tar -C /mnt/ost</screen></para>
651 <para>Recreate the OST configuration files, if unavailable. </para>
652 <para>Follow the procedure in
653 <xref linkend="dbdoclet.repair_ost_lastid"/> to recreate the LAST_ID
654 file for this OST index. The <literal>last_rcvd</literal> file
655 will be recreated when the OST is first mounted using the default
656 parameters, which are normally correct for all file systems. The
657 <literal>CONFIGS/mountdata</literal> file is created by
658 <literal>mkfs.lustre</literal> at format time, but has flags set
659 that request it to register itself with the MGS. It is possible to
660 copy the flags from another working OST (which should be the same):
661 <screen>oss1# debugfs -c -R "dump CONFIGS/mountdata /tmp" <replaceable>/dev/other_osdev</replaceable>
662 oss1# scp /tmp/mountdata oss0:/tmp/mountdata
663 oss0# dd if=/tmp/mountdata of=/mnt/ost/CONFIGS/mountdata bs=4 count=1 seek=5 skip=5 conv=notrunc</screen></para>
666 <para> Unmount the OST file system.
667 <screen>oss# umount /mnt/ost</screen>
672 <section xml:id="section_ucf_qgt_tl">
674 <primary>maintenance</primary>
675 <secondary>reintroducing an OSTs</secondary>
676 </indexterm>Returning a Deactivated OST to Service</title>
677 <para>If the OST was permanently deactivated, it needs to be
678 reactivated in the MGS configuration.
679 <screen>mgs# lctl conf_param <replaceable>ost_name</replaceable>.osc.active=1</screen>
680 If the OST was temporarily deactivated, it needs to be reactivated on
682 <screen>mds# lctl set_param osp.<replaceable>fsname</replaceable>-OST<replaceable>number</replaceable>-*.active=1
683 client# lctl set_param osc.<replaceable>fsname</replaceable>-OST<replaceable>number</replaceable>-*.active=1</screen></para>
686 <section xml:id="dbdoclet.50438199_77819">
687 <title><indexterm><primary>maintenance</primary><secondary>aborting recovery</secondary></indexterm>
688 <indexterm><primary>backup</primary><secondary>aborting recovery</secondary></indexterm>
689 Aborting Recovery</title>
690 <para>You can abort recovery with either the <literal>lctl</literal> utility or by mounting the target with the <literal>abort_recov</literal> option (<literal>mount -o abort_recov</literal>). When starting a target, run: <screen>mds# mount -t lustre -L <replaceable>mdt_name</replaceable> -o abort_recov <replaceable>/mount_point</replaceable></screen></para>
692 <para>The recovery process is blocked until all OSTs are available. </para>
695 <section xml:id="dbdoclet.50438199_12607">
696 <title><indexterm><primary>maintenance</primary><secondary>identifying OST host</secondary></indexterm>
697 Determining Which Machine is Serving an OST </title>
698 <para>In the course of administering a Lustre file system, you may need to determine which
699 machine is serving a specific OST. It is not as simple as identifying the machine’s IP
700 address, as IP is only one of several networking protocols that the Lustre software uses and,
701 as such, LNet does not use IP addresses as node identifiers, but NIDs instead. To identify the
702 NID that is serving a specific OST, run one of the following commands on a client (you do not
703 need to be a root user):
704 <screen>client$ lctl get_param osc.<replaceable>fsname</replaceable>-OST<replaceable>number</replaceable>*.ost_conn_uuid</screen>
706 <screen>client$ lctl get_param osc.*-OST0000*.ost_conn_uuid
707 osc.testfs-OST0000-osc-f1579000.ost_conn_uuid=192.168.20.1@tcp</screen>-
709 <screen>client$ lctl get_param osc.*.ost_conn_uuid
710 osc.testfs-OST0000-osc-f1579000.ost_conn_uuid=192.168.20.1@tcp
711 osc.testfs-OST0001-osc-f1579000.ost_conn_uuid=192.168.20.1@tcp
712 osc.testfs-OST0002-osc-f1579000.ost_conn_uuid=192.168.20.1@tcp
713 osc.testfs-OST0003-osc-f1579000.ost_conn_uuid=192.168.20.1@tcp
714 osc.testfs-OST0004-osc-f1579000.ost_conn_uuid=192.168.20.1@tcp</screen></para>
716 <section xml:id="dbdoclet.50438199_62333">
717 <title><indexterm><primary>maintenance</primary><secondary>changing failover node address</secondary></indexterm>
718 Changing the Address of a Failover Node</title>
719 <para>To change the address of a failover node (e.g, to use node X instead of node Y), run
720 this command on the OSS/OST partition (depending on which option was used to originally
722 <screen>oss# tunefs.lustre --erase-params --servicenode=<replaceable>NID</replaceable> <replaceable>/dev/ost_device</replaceable></screen>
724 <screen>oss# tunefs.lustre --erase-params --failnode=<replaceable>NID</replaceable> <replaceable>/dev/ost_device</replaceable></screen>
725 For more information about the <literal>--servicenode</literal> and
726 <literal>--failnode</literal> options, see <xref xmlns:xlink="http://www.w3.org/1999/xlink"
727 linkend="configuringfailover"/>.</para>
729 <section xml:id="dbdoclet.50438199_62545">
730 <title><indexterm><primary>maintenance</primary><secondary>separate a
731 combined MGS/MDT</secondary></indexterm>
732 Separate a combined MGS/MDT</title>
733 <para>These instructions assume the MGS node will be the same as the MDS
734 node. For instructions on how to move MGS to a different node, see
735 <xref linkend="dbdoclet.changingservernid"/>.</para>
736 <para>These instructions are for doing the split without shutting down
737 other servers and clients.</para>
740 <para>Stop the MDS.</para>
741 <para>Unmount the MDT</para>
742 <screen>umount -f <replaceable>/dev/mdt_device</replaceable> </screen>
745 <para>Create the MGS.</para>
746 <screen>mds# mkfs.lustre --mgs --device-size=<replaceable>size</replaceable> <replaceable>/dev/mgs_device</replaceable></screen>
749 <para>Copy the configuration data from MDT disk to the new MGS disk.</para>
750 <screen>mds# mount -t ldiskfs -o ro <replaceable>/dev/mdt_device</replaceable> <replaceable>/mdt_mount_point</replaceable></screen>
751 <screen>mds# mount -t ldiskfs -o rw <replaceable>/dev/mgs_device</replaceable> <replaceable>/mgs_mount_point</replaceable> </screen>
752 <screen>mds# cp -r <replaceable>/mdt_mount_point</replaceable>/CONFIGS/<replaceable>filesystem_name</replaceable>-* <replaceable>/mgs_mount_point</replaceable>/CONFIGS/. </screen>
753 <screen>mds# umount <replaceable>/mgs_mount_point</replaceable></screen>
754 <screen>mds# umount <replaceable>/mdt_mount_point</replaceable></screen>
755 <para>See <xref linkend="dbdoclet.50438199_54623"/> for alternative method.</para>
758 <para>Start the MGS.</para>
759 <screen>mgs# mount -t lustre <replaceable>/dev/mgs_device</replaceable> <replaceable>/mgs_mount_point</replaceable></screen>
760 <para>Check to make sure it knows about all your file system</para>
761 <screen>mgs:/root# lctl get_param mgs.MGS.filesystems</screen>
764 <para>Remove the MGS option from the MDT, and set the new MGS nid.</para>
765 <screen>mds# tunefs.lustre --nomgs --mgsnode=<replaceable>new_mgs_nid</replaceable> <replaceable>/dev/mdt-device</replaceable></screen>
768 <para>Start the MDT.</para>
769 <screen>mds# mount -t lustre <replaceable>/dev/mdt_device /mdt_mount_point</replaceable></screen>
770 <para>Check to make sure the MGS configuration looks right:</para>
771 <screen>mgs# lctl get_param mgs.MGS.live.<replaceable>filesystem_name</replaceable></screen>