Zabbix Disabling Journaling on ext4
Zabbix server 2.2.3 (VM on VMWare 5.5) OS: CentOS release 6.5 (Final)
CPU utilization metrics:

Task: Disable journaling on ext4.
HDD:
[root@v11 /]# df -T
Filesystem Type 1K-blocks Used Available Use% Mounted on
/dev/mapper/vg_zab-lv_root
ext4 14356848 6182300 7445256 46% /
tmpfs tmpfs 961212 0 961212 0% /dev/shm
/dev/sda1 ext4 495844 54392 415852 12% /boot
Check filesystem characteristics:
[root@v11 /]# dumpe2fs /dev/mapper/vg_zab-lv_root |grep journal
dumpe2fs 1.41.12 (17-May-2010)
Filesystem features: has_journal ext_attr resize_inode dir_index filetype needs_recovery extent flex_bg sparse_super large_file huge_file uninit_bg dir_nlink extra_isize
Journal features: journal_incompat_revoke
Executed on a live system:
[root@v11 /]# tune2fs -o journal_data_writeback /dev/mapper/vg_zab-lv_root
tune2fs 1.41.12 (17-May-2010)
Verifying:
[root@v11 /]# tune2fs -l /dev/mapper/vg_zab-lv_root | grep writeback
Default mount options: journal_data_writeback user_xattr acl
Rebooted the server. Checked. No expected effect.
Shut down the server. Booted from a LiveCD that has LVM support (GParted LiveCD 0.19.1-1). Activated LVM:
Completely disabled journaling on the partition:
user@debian:~$ sudo tune2fs -O ^has_journal /dev/mapper/vg_zab-lv_root
tune2fs 1.42.11 (09-Jul-2014)
Rebooted the server. Checked iotop — the system journaling process is no longer in the top lines. After a brief analysis, I concluded that the regular messages correspond to the hourly Housekeeper process, which heavily loads the database.
Conclusions: 1. No effect from disabling journaling on this configuration. 2. With this number of hosts, I don't see this event as critical.