My computer suffers from sudden and continous hard drive load strokes. Sometimes it lasts for a few minutes and hence work is impossible because everything goes very slow.
\nI’m trying to locate the cause because it makes me nervous :)<\/p>\n
Today I found one of the causes. It’s\u00a0updatedb.mlocate<\/i> script which is responsible for scanning hard drives and build locate (a location of files) database. But due to some NTFS driver limitations that poor thing thinks that NTFS share is always new and needs to be rescan at every scheduled scan.<\/p>\n
How to make sure that you have the same problem? When hard drive starts to choke check all IO processes with<\/p>\n
sudo iotop<\/i><\/p>\n
and look for\u00a0updatedb.mlocate <\/i>or\u00a0mount.ntfs<\/i>\u00a0processes with high IO load. If this is the case try to fix it.
\nThe solution is to point NTFS shares as non-scannable. To do this edit<\/p>\n
\/etc\/updatedb.conf<\/i><\/p>\n
and add your NTFS mount paths to\u00a0PRUNEPATHS list. Although NTFS is listed in PRUNEFS entry that dummy mlocate<\/i> script still would scan NTFS shares.<\/p>\n