High CPU usage - Have a log file
-
Hello (sorry for my bad english), I'm having a trouble with my pfsense, I build it myself with my hardware that was recommended here and it was working great for like a year, I have site 2 site to my other pfsense, pfblocker, snort, its great! (pfsense is updated to the latest version, same for the packages)
but I'm having a weird thing that is happening randomly recently and I hope someone could help me figuring out what seems to be the problem, I'm losing the internet.. I cannot at that time even connect to the router itself and from what I could find its a high cpu usage, restarting the router fix the problem but it comes back after few days.
I have a log of the last time it happen, the problem starts at Jan 21 23:55:54,
pfsense-log.txt -
Jan 21 23:55:54 kernel igb0: link state changed to DOWN
The igb0 NIC was disconnected. Was are you using that NIC for? What is it connected to?
Steve
-
yes, I was using it for like a year already, its connected with an adapter to the motherboard,
I still can ping to the router when the problem happens but cannot connect because of the high cpu,
the nic can cause that? high cpu usage? -
The NIC was not disconnected from the system, that would be much worse. The igb0 NIC lost link, the Ethernet connection was lost. What's on the other end of that cable? What is the condition of the cable?
Steve
-
@stephenw10 well that would be the isp modem, the igb0 is my wan, keep in mind that if i physically disconnect the modem, this cpu usage don't happen
-
How high is the CPU usage exactly? The logs show a bunch of churn as the ppp daemon has to exit and restart. I assume that is running on igb0.
Something is triggering that message.I would swap out the cables if you have not already. Then I would switch the WAN to a different NIC and see if the fault follows it.
Steve
-
@stephenw10 CPU usage is on 100%
Thank you, I'll try that and update you :) -
Check the output of
top -aSH
at the command line to see what is using it. -
@stephenw10 I have an update about the issue, well first I did what you told me and the CPU usage was because of this:
93252 unbound 52 0 347M 339M kqread 0 0:01 77.69% /usr/local/sbin/unbound -c /var/unbound/
93252 unbound 52 0 347M 339M kqread 1 0:30 74.07% /usr/local/sbin/unbound -c /var/unbound/which is the dns resolver as far as i know but its not what caused the issue,
this is the NIC I use (Intel i350-T4):
https://www.ebay.com/c/919241189and yes I got it from this guy on ebay, now I know its a counterfeit card and it was working perfectly fine until now,
I found a topic about this cards here:
https://forums.servethehome.com/index.php?threads/comparison-intel-i350-t4-genuine-vs-fake.6917/
people report that after a few months they start to fail and I had it over a year and a half, I'm almost sure that the issue is because of the NIC,
I checked it and found some corrosion on it, I'm adding a photo of it -
Ouch!
Hard to see why that would have caused such problems for Unbound though whilst other traffic was passing. If it loses connection entirely Unbound might use significant CPU trying to connect, though that still looks high.If the card is failing though it could fail in interesting ways, almost anything could happen!
Steve