Hi Kurt, this seems to be your month for lustre adventures :/

Best I can tell this message means you have a directory with ~16M entries. That is a hard limit in older versions of ext3-based lustre (http://ext2.sourceforge.net/2005-ols/paper-html/node3.html)

According to LU-1365 (https://jira.hpdd.intel.com/browse/LU-1365). If you have a MDT formatted under an older version of lustre (specifically used old version mkfs.lustre) you get this message because at the time of the LU ticket, the e2fs userspace tools (ie. fsck) could not handle LARGEDIR feature.

I think you can ignore this warning if you are willing to live with the directory 16M entry limit. I would _guess_ this is fixed in newer version of e2fsprogs. However there may not be a way to add LARGEDIR feature without re-formatting the MDT volume.

regards,
chris hunter

Date: Thu, 28 Jan 2016 16:15:41 -0500 (EST)
From: Kurt Strosahl<stros...@jlab.org>
To:lustre-discuss@lists.lustre.org
Subject: [lustre-discuss] Large directory Feature in lustre 2.5.3
Message-ID:<189341774.1403835.1454015741354.javamail.zim...@jlab.org>
Content-Type: text/plain; charset=utf-8

Hello,

    Recently this message appeard in the metadata system on our lustre 2.5.3 
file system.  Is there a way to enable this feature in 2.5.3?

Jan 28 10:53:56 scmds14a kernel: LDISKFS-fs warning (device dm-1): 
ldiskfs_dx_add_entry: Directory (ino: 111675137) index full, reach max htree 
level :2
Jan 28 10:53:56 scmds14a kernel: LDISKFS-fs warning (device dm-1): 
ldiskfs_dx_add_entry: Large directory feature isnot enabled on this filesystem

w/r,
Kurt J. Strosahl
System Administrator
Scientific Computing Group, Thomas Jefferson National Accelerator Facility
_______________________________________________
lustre-discuss mailing list
lustre-discuss@lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org

Reply via email to