Firstly I want to congratulate you for the excellent work that you realized this year, it is very appreciated! I hope that the year 2019 will be as good in features :-)
This topic about a small issue/overload increase on a VPS node at midnight, after checking it would appear to be the CRON defined in /etc/cron.d/lfd-cron
Code: Select allThis restart is really quick but it's look make some checksum in /usr/bin/, /usr/sbin/ and /bin/:
[root@cPanel ~]# cat /etc/cron.d/lfd-cron SHELL=/bin/sh 0 0 * * * root /usr/sbin/csf --lfd restart > /dev/null 2>&1
Code: Select all=> 209863 files are checked during the CSF restart X guest hosts on the same node.
/usr/bin/md5sum --check /var/lib/csf/csf.tempint lfd - messenger csf... lfd - sleeping ----- [root@cPanel ~]# cat /var/lib/csf/csf.tempint | wc -c 209863
We like the features "System Integrity Checking" but in this case we prefer to launch it not in the same to spread the load/iowait :-)
We have tried to randomize the call to this restart like this:
Code: Select allHowever, this file is overwritten with each CSF update:[/b]
sed -i "s/0 0 * * */$(( $RANDOM % 10 )) $(( $RANDOM % 10 )) * * */" /etc/cron.d/lfd-cron
Code: Select allFirst, what is the usefulness of the restart lfd if the process is already started?
[root@cPanel ~]# stat /etc/cron.d/lfd-cron File: ‘/etc/cron.d/lfd-cron’ Size: 74 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 26494092 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2018-12-27 23:17:11.704332671 +0100 Modify: 2018-12-05 06:42:04.131057712 +0100 Change: 2018-12-05 06:42:05.118033505 +0100 Birth: -
Second, could we have a configuration that would launch this CRON at different times of the day for each guest host?
A big thank you for reading this message/request.