Mdadm is my beloved friend but something that irritates a lot – periodic inspections and resink the security health of RAID array- for example, there are data in a bad sector-and, which in turn crushed by the machine to the IO. Basically after I found čoplene guilty – branches that start usually around 1 pm every Sunday night. The idea is clear – sure that the array is in great condition and there are no dramas with the information. This is good but it seems many weekly, that's why it prekonfigurirah to r″nva of the first date of the month.
For Redhat-based derivatives path of Krone's /etc/cron.d/raid-check. For Debian based distroci's way /etc/cron.d/mdadm. Kronovete in turn called bash scripts /usr/sbin/raid-check for CentOS etc and /usr/share/mdadm/checkarray for Debian and friends. Parameters to the scripts are taken from /etc/sysconfig/raid-check or, respectively, /etc/default/mdadm where you may be banned entirely and check, that is not a very smart idea.
Some developers just won't learn to write literate in RFC never. I noticed multiple errror_log files that had accumulated a huge number of retarded warning and notice for failure to comply with the standard PHP. In General, it is difficult to explain to the user, that code which is put is bad and needs to be fixed. In the General case, I have noticed that users do not excite them error-log after running the code. In General, a radical approach to stop entirely error_log files and who wants to let them go, but as a whole would create discomfort for a lot of users. So strengthen the approach 2 – adminski super powers or 1 bash line. Search for files named error_log with size more than 5MB (Here I leave my value a although 1MB is more than enough) and deleting them weekly. This effect is achieved with clutter find
find /home/ -name error_log -size +5M -type f -delete
It remains only to crash into the Crown to perform once a week and we have a lot of persistentno solution. In my case it seems OK in 1 o'clock in the evening every Sunday.