https://jira.whamcloud.com This file is an XML representation of an issue en-us 9.4.14 940014 05-12-2023 https://jira.whamcloud.com/browse/LU-4569 Lustre <p>The posix copytool imports files that are designated by their path in the backend, despite being itself the only one to know this path.</p> <p>We propose a slight improvement to make the posix CT also accept a FID as import source parameter (from which the path can be rebuilt). This would allow easy "undeletion" of archived files:</p> <p>lhsmtool_posix --import &lt;fid&gt; /mnt/lustre/undelete ...</p> LU-4569

Posix copytool import a file using only its FID

Improvement Minor Resolved Fixed Bruno Faccini Henri Doreau HSM patch Fri, 31 Jan 2014 08:47:42 +0000 Wed, 11 Jun 2014 22:11:43 +0000 Tue, 20 May 2014 13:26:07 +0000 Lustre 2.6.0 Lustre 2.5.2 0 8 <p>Patch at <a href="http://review.whamcloud.com/9075" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/9075</a></p> <p>Bruno</p> <p>Could you please take care of this patch?</p> <p>Thanks</p> <p>Peter</p> <p>The only few auto-tests errors are unrelated to this patch. Looks good to me.</p> <p>Any update on this?</p> <p>Hello Henry, reviews are in progress so I think your patch should land soon now!</p> <p>Landed for 2.6</p> <p>Patch for b2_5 at <a href="http://review.whamcloud.com/#/c/10132/" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/#/c/10132/</a></p> <p>This patch missed the following use case:<br/> 1) create/archive a file1 with FID1<br/> 2) import a new file2 from FID1<br/> 3) restore file2<br/> 4) file1 and file2 share the same archive in backend so it will be corrupted after next archive of file1 or file2</p> <p>solution:<br/> (2) must be forbidden if FID1 exists in Lustre </p> <p>Thanks JC. I will submit a fix for master and an updated patch for 2.5.</p> <p>Patch for master: <a href="http://review.whamcloud.com/10185" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/10185</a><br/> The problem was already present when importing by path, the patch aims at fixing both cases.</p> <p>The check on import by path breaks sanity-hsm. Patchset #2 only performs the existence check when importing FIDs (more error-prone maybe?). Also, the extra access() call has a non-negligible impact, that might be undesirable when importing large directory tree which is typically what import-by-path would be useful for.</p> <p>The problem is that sanity-hsm directly copies files to the archive without following the naming scheme used by the posix copytool. An option would be to copy those files into lustre, archiving them and removing them from lustre, but this is not totally equivalent (it involves many more components). Let me know if some solution is preferred.</p> <p>Both patches landed to 2.6; will be landed to b2_5 separately</p> Development Rank 1|hzwdyn: Rank (Obsolete) 12471