RE: QTOBDNS - disk busy percent

I had an issue back in 2015 (V7R1) where the a system task consuming 30% of CPU - EL-ERRLOG 29.6
Not exactly your issue.
Thousands of 1800 0061 VLOGs were created.
MF60070 and MF60348 resolved the issue.
problem gets triggered by
erroneous PALs being encountered, then the PAL size counter gets an incorrect value.
Once that size counter goes invalid, the VLOGs start.

Check your SST logs, PAL, LIC logs, etc.

Paul


-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxxxxxxxx] On Behalf Of T. Adair
Sent: Friday, May 17, 2019 3:34 PM
To: midrange-l@xxxxxxxxxxxxxxxxxx
Subject: QTOBDNS - disk busy percent

One night this week our System i went from averaging 0-3% busy on all
disks (which is normal) to 30-50% busy and stayed there. This destroyed
our response time, making the system barely usable.

For 2½ days we tried everything we could think of, all to no avail. Our
CPU utilization was fine, our disk space was fine, there were no
'runaway' jobs, etc. Our disks were simply getting hammered. We
brought in our BP and he adjusted our memory allocation, moving memory
from two unused partitions, and that helped tremendously. Our page
faulting was within normal expectations, so why would additional memory
help?

But here's my real question. While searching for the culprit, I ran
across a system (QTCP) job, QTOBDNS. I was surprised at how much CPU it
was taking so I checked its job log. There were hundreds (if not
thousands) of entries that really concern me. Example...:
mixpanel.com
aniview.com
d.turn.com
brealtime.com
akamaiedge.com
rubiconproject.com
googleapis.com
taboola.com
(you get the idea)

I have a basic understanding of what a DNS does but this really seems
strange. Could this job be the source of our original problem? And
here's the question I hate to ask, but need to: is it possible we've
been hacked?

We're currently on 7.2. And no, we're not up-to-date on PTFs.

Thanks in advance for any thoughts on this.

~TA~