performance – Disk saturated by mariadb on CentOS 7, blocks everything

For the past few days I have been having a problem with a webserver running CentOS 7. Every 13-14 minutes some operation (which I identified with mariadb using iostat) seems to completely saturate the disk IO operations, blocking all other applications. After about 3 minutes everything returns to normal.

In /var/log/messages I see:

Sep 14 09:35:42 MY-VM kernel: mptscsih: ioc0: attempting task abort! (sc=ffff9aefd1e76d80)
Sep 14 09:35:42 MY-VM kernel: sd 0:0:0:0: [sda] CDB: Write(10) 2a 00 2a 8b 1b 18 00 00 10 00
Sep 14 09:35:42 MY-VM kernel: mptscsih: ioc0: task abort: SUCCESS (rv=2002) (sc=ffff9aefd1e76d80)
Sep 14 09:35:42 MY-VM kernel: mptscsih: ioc0: attempting task abort! (sc=ffff9aefd1e741c0)
Sep 14 09:35:42 MY-VM kernel: sd 0:0:0:0: [sda] CDB: Write(10) 2a 00 0f 37 0a 58 00 00 08 00
Sep 14 09:35:42 MY-VM kernel: mptscsih: ioc0: task abort: SUCCESS (rv=2002) (sc=ffff9aefd1e741c0)
Sep 14 09:35:42 MY-VM kernel: mptscsih: ioc0: attempting task abort! (sc=ffff9aefd1e74700)
Sep 14 09:35:42 MY-VM kernel: sd 0:0:0:0: [sda] CDB: Write(10) 2a 00 0f 37 0a 38 00 00 10 00
Sep 14 09:35:42 MY-VM kernel: mptscsih: ioc0: task abort: SUCCESS (rv=2002) (sc=ffff9aefd1e74700)

I’ve been monitoring the queries, but I can’t find anything significant.