arp: attempts to modify permanent entry

[HolyK]

Well, i guess i saw something similar here but since search is somehow "odd" here then i will start over ...

I am running: FreeNAS-9.2.0-RELEASE-x64 (ab098f4)

I have two jails running ...
- Pluginjail with Transmission (epair0b)
- Portjail with ZNC bouncer (epair1b)

Time to time i am facing a situation that one of the FANs (95% that it's a CPU fan) get crazy and running on max speed which is noisy and annoying, especially when working or watching a movie. That noise is not a question here since it is a BOXed FAN and i was expecting that one day it will start doing annoying noise. The question is WHY it suddenly starts rotating at max (?) speed every day.

First i thought that it's because of the hot days here but i was working from home several days in a row and find out that it's starting every day at the evening. During the night/forenoon it's perfectly quiet.

I verified that i don't have any scrub or periodical job running on daily basis which could somehow causing this. So i went through logs and find following message spamming logfile a LOT.

Thing is that i did not found that MAC anywhere on my network (checked currently connected devices, NAS, jails, DHCP server). I also tried to sniff networkflow on my Mikrotik router but there is nothing with that MAC. 192.168.xxx.yyy is IP of my Router (Gateway)

Any ideas what should i do with that ? From where its coming?

EDIT: I just find out that the MAC address belongs to one of the ETH ports on Router on which Wireless AP is connected in bridge mode (so via LAN port, no routing or whatever, just a bridge). Anyway i still have no idea what should i do with that...

Code:

Jun 15 18:29:45 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair0b
Jun 15 18:29:46 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair1b
Jun 15 18:30:06 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair0b
Jun 15 18:30:09 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair1b
Jun 15 18:30:27 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair0b
Jun 15 18:30:42 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair1b
Jun 15 18:31:12 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair0b
Jun 15 18:31:32 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair1b
Jun 15 18:31:58 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair0b
Jun 15 18:32:24 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair1b
Jun 15 18:32:43 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair0b
Jun 15 18:33:27 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair1b
Jun 15 18:33:29 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair0b
Jun 15 18:34:13 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair1b
Jun 15 18:34:14 HolyNAS kernel: arp: zz.zz.zz.zz.zz.zz attempts to modify permanent entry for 192.168.xxx.yyy on epair0b

 

[DrKK]

At the times that fan is running at max speed, do you have a CPU temperature issue?

Code:

[root@freenas] ~# sysctl -a | grep temperature
hw.acpi.thermal.tz0.temperature: 27.8C
hw.acpi.thermal.tz1.temperature: 29.8C
dev.cpu.0.temperature: 31.0C
dev.cpu.1.temperature: 32.0C

As you can see, my CPU is running at about ~30C, which is fine, and should have the fan on minimum.

Next time your fan goes crazy, issue the above command from the FreeNAS command-line, and tell us what your temperatures are.

 

[HolyK]

Yea, probably temp issue. Right now it is "quiet", temps are fine EDIT: No, they are not (although my ass is melting on the chair...) and there is no "spam" of that message.

[root@HolyNAS] /var/log# sysctl -a | grep temperature
hw.acpi.thermal.tz0.temperature: 27.8C
hw.acpi.thermal.tz1.temperature: 29.8C
dev.cpu.0.temperature: 51.0C
dev.cpu.1.temperature: 51.0C
dev.cpu.2.temperature: 48.0C
dev.cpu.3.temperature: 48.0C

I will re-check the temp next time. I am almost 100% sure that its related ...

 

[DrKK]

No sir,

The temps are *NOT* fine. If your computer is sitting there idling at 50C, then the temps are most certainly *NOT* fine.

You have a thermal issue. Most likely culprit is either bad thermal paste, or the fan/heatsink are not properly attached.

 

[DrKK]

I'm assuming, of course, you have a proper desktop/server CPU. Temps like these would be not out of the ordinary for certain kinds of old-school Intel atoms, or AMD E-350 or E-450.

 

[HolyK]

Well it's i3 4130 which normally should have like ~37ºC in CPU and ~34ºC in cores based on the discussions google provided and the max temp is 72ºC. I admit that the 50ºC on core is a bit higher but consider the fact that it has low profile BOXed FAN (which is piece of crap as usual) and system is not complete idle (but neither so loaded). Also we're having a hot days here (like i said above), so increased temperature is expected. If the air around me get just a little bit higher i will i will start cooking myself in my own juice and i will need a shovel to dig myself out of here since the air is heavy like a pig...

Anyway i will get new CPU FAN this week because i am afraid that the CPU will melt down, true summer did not yet started ...

 

[cyberjock]

50C is cause for concern in my book for a system that isn't heavily loaded.

 

[DrKK]

50C is not correct for an i3-4130, even in a hot room, unless it is under full load. You can either trust us, or not trust us. Ask yourself if we sound like we know what we're talking about. :)

You have one of the following 3 problems, with probability 99.14%:

1) Your CPU is under full load all the time for some reason,
2) You have a massive air-flow problem and/or fan problem in the case (that being said, I've never once in my life seen a case airflow problem that would leave an otherwise correctly functioning CPU heatsink/fan with a 50C temp),
3) Your heatsink is not correctly clamped on to the CPU case, and/or, the thermal interface between the two is not performing correctly (either because it's gone, shitty, or incorrectly applied).

Based upon the symptoms you've said, I doubt it's #1 or #2. I am going with #3.

 

[Ericloewe]

50C is not correct for an i3-4130, even in a hot room, unless it is under full load. You can either trust us, or not trust us. Ask yourself if we sound like we know what we're talking about. :)

You have one of the following 3 problems, with probability 99.14%:

1) Your CPU is under full load all the time for some reason,
2) You have a massive air-flow problem and/or fan problem in the case (that being said, I've never once in my life seen a case airflow problem that would leave an otherwise correctly functioning CPU heatsink/fan with a 50C temp),
3) Your heatsink is not correctly clamped on to the CPU case, and/or, the thermal interface between the two is not performing correctly (either because it's gone, shitty, or incorrectly applied).

Based upon the symptoms you've said, I doubt it's #1 or #2. I am going with #3.

I'd bet on 3). Intel's stock coolers are a pain to properly install.

 

[cyberjock]

I've never had a problem installing an Intel cooler. I've always gotten it right, and always on the first try. But, being that I'm an engineering type and very interested in things such as heat transfer I'm probably not a good point of reference. ;)

I've also never removed a cooler for any reason except to upgrade a CPU. Never believed the story that you should replace your thermal paste regularly.

 

[HolyK]

1] Nope
2] Me neither
3] Well, system is running almost year without any major hiccup and i find old session log with temps like ~31ºC. So my guess is that the Intel FAN is just fcked up. Either one of the plastic holders popped out so the heatsink is not aligned properly or that FAN is simple screwed.

I'll take a look on it and let you know...

 

[DrKK]

3] Well, system is running almost year without any major hiccup and i find old session log with temps like ~31ºC. So my guess is that the Intel FAN is just fcked up. Either one of the plastic holders popped out so the heatsink is not aligned properly or that FAN is simple screwed.

I'm almost positive you'll find it's one of those stupid push pins. It's easy to only 85% push them in initially, then over time, the thermal interface isn't snug. As I said, and as I think you now agree, 50C is not possible for an IDLE temperature. Something must be wrong, physically, with the cooling solution, sir.

 

[HolyK]

In the meanwhile any ideas what can i do with that logspam ?

 

[c32767a]

In the meanwhile any ideas what can i do with that logspam ?

Probably just ignore it, or disable whatever device is ARPing for your NAS' IP addresses.
The message comes from the BSD IP stack and is intended to alert the sysadmin to what the BSD IP stack considers to be a Bad Thing. It may indicate an actual problem on your network, or it may be superfluous.
Without more troubleshooting and knowledge of the layout of your network, it's hard to say exactly what's happening, but you see this a lot when things are doing gratuitous ARPs, or you have duplicate IPs, or other protocols are running on the network that do things BSD doesn't expect. (see the mesh router discussion from a few weeks ago.)

 

[HolyK]

Ok i found it (i guess). I had Proxy-arp enabled on both master-switch and AP ports so disabled on the second one.

I am not hearing that fan sound anymore or maybe i just can't hear it because GPU FAN from Desktop is much louder ... Divinity is giving hard times to my old BFG 280gtx (R.I.P. BFG comp.) ^^

Anyway CPU temp is still quite high so i have Noctua already on my table and will switch cooler this week :)

Sent from my Motorola Photon Q

 

[HolyK]

[root@HolyNAS] ~# sysctl -a | grep temperature
hw.acpi.thermal.tz0.temperature: 27.8C
hw.acpi.thermal.tz1.temperature: 29.8C
dev.cpu.0.temperature: 36.0C
dev.cpu.1.temperature: 36.0C
dev.cpu.2.temperature: 36.0C
dev.cpu.3.temperature: 36.0C

Better i guess? It's like 15 minutes after NAS start and today is an awful weather here, so not a crazy-hot day at all. So i guess temperature will raise a bit. And please don't say that the temp is too much ... Noctua NH-U12S replaced that BOXed crap ^^

 

[Whattteva]

Not sure what temperature the tz's are, but 36 C for CPU is really good considering the upper limit for most CPU's are typically in the 70's if I remember correctly. Of course, you don't want to hover at 60+ for long.