Unraid Server NTPd e1000 nic resets :waking up resolver.

[thestraycat]

Been monitoring my logs all day and there full of the following:

 

active_time=60 secs

Nov 4 12:03:28 MediaServer ntpd[1832]: 193.150.34.2 local addr 192.168.1.125 -> <null>

Nov 4 12:03:31 MediaServer kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx

Nov 4 12:03:32 MediaServer ntpd[1832]: Listen normally on 236 eth0 192.168.1.125:123

Nov 4 12:03:32 MediaServer ntpd[1832]: 193.150.34.2 local addr 172.27.239.1 -> 192.168.1.125

Nov 4 12:03:32 MediaServer ntpd[1832]: new interface(s) found: waking up resolver

Nov 4 12:11:51 MediaServer kernel: e1000e 0000:00:19.0 eth0: Reset adapter unexpectedly

Nov 4 12:11:52 MediaServer ntpd[1832]: Deleting interface #236 eth0, 192.168.1.125#123, interface stats: received=4, sent=7, dropped=0, active_time=500 secs

Nov 4 12:11:52 MediaServer ntpd[1832]: 193.150.34.2 local addr 192.168.1.125 -> <null>

Nov 4 12:11:55 MediaServer kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx

Nov 4 12:11:56 MediaServer ntpd[1832]: Listen normally on 237 eth0 192.168.1.125:123

Nov 4 12:11:56 MediaServer ntpd[1832]: 193.150.34.2 local addr 172.27.239.1 -> 192.168.1.125

Nov 4 12:11:56 MediaServer ntpd[1832]: new interface(s) found: waking up resolver

Nov 4 12:28:50 MediaServer kernel: e1000e 0000:00:19.0 eth0: Reset adapter unexpectedly

Nov 4 12:28:51 MediaServer ntpd[1832]: Deleting interface #237 eth0, 192.168.1.125#123, interface stats: received=9, sent=11, dropped=0, active_time=1015 secs

Nov 4 12:28:51 MediaServer ntpd[1832]: 193.150.34.2 local addr 192.168.1.125 -> <null>

Nov 4 12:28:54 MediaServer kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx

Nov 4 12:28:55 MediaServer ntpd[1832]: Listen normally on 238 eth0 192.168.1.125:123

Nov 4 12:28:55 MediaServer ntpd[1832]: 193.150.34.2 local addr 172.27.239.1 -> 192.168.1.125

Nov 4 12:28:55 MediaServer ntpd[1832]: new interface(s) found: waking up resolver

Nov 4 12:29:49 MediaServer kernel: e1000e 0000:00:19.0 eth0: Reset adapter unexpectedly

Nov 4 12:29:50 MediaServer ntpd[1832]: Deleting interface #238 eth0, 192.168.1.125#123, interface stats: received=0, sent=1, dropped=0, active_time=55 secs

Nov 4 12:29:50 MediaServer ntpd[1832]: 193.150.34.2 local addr 192.168.1.125 -> <null>

Nov 4 12:29:53 MediaServer kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx

Nov 4 12:29:54 MediaServer ntpd[1832]: Listen normally on 239 eth0 192.168.1.125:123

Nov 4 12:29:54 MediaServer ntpd[1832]: 193.150.34.2 local addr 172.27.239.1 -> 192.168.1.125

Nov 4 12:29:54 MediaServer ntpd[1832]: new interface(s) found: waking up resolver

Nov 4 12:32:23 MediaServer kernel: e1000e 0000:00:19.0 eth0: Reset adapter unexpectedly

Nov 4 12:32:24 MediaServer ntpd[1832]: Deleting interface #239 eth0, 192.168.1.125#123, interface stats: received=0, sent=3, dropped=0, active_time=150 secs

Nov 4 12:32:24 MediaServer ntpd[1832]: 193.150.34.2 local addr 192.168.1.125 -> <null>

Nov 4 12:32:27 MediaServer kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx

Nov 4 12:32:28 MediaServer ntpd[1832]: Listen normally on 240 eth0 192.168.1.125:123

Nov 4 12:32:28 MediaServer ntpd[1832]: 193.150.34.2 local addr 172.27.239.1 -> 192.168.1.125

Nov 4 12:32:28 MediaServer ntpd[1832]: new interface(s) found: waking up resolver

 

Am i right in thinking that the NTP daemon is crashing my nic's and which are then resetting themselves and reregistering as a new NIC in unraid every time?? I've turned off NTP for now.... To see if it calms down and stabilises it came on from installing a few dockers within 24hrs. Now the server needs a daily reboot and takes 15-20 mins to come up. Can post logs/diag etc but thought i'd just test the water first in case anyone had seen the issue?

 

Anyone seen this behaviour?

 

 

[John_M]

It looks as though your NIC keeps resetting but I think the ntpd log entries are the effect, not the cause.

 

[thestraycat]

Thanks John - Any ideas what may cause it? I would upload the diag but it's just hung and is unresponsive. i'm assuming when i hard off it it will have lost all of it's logs... Is there an easy way to forward the log to a remote pc on the network anyone?

[John_M]

It could be hardware related. Is the NIC on a card? You could try reseating it. If that doesn't work you could try it in a different slot or replace it. I don't think it's cable related but it wouldn't hurt to try a new one.

[John_M]

If you can telnet or ssh in you can type

 

diagnostics

 

and they'll be saved to your flash device.

 

EDIT: If you can't telnet or ssh in you could plug in a keyboard and just type.

 

[John_M]

I just Googled your motherboard. It looks as though it has two on-board NICs. If this is one of them you could try using the other one instead.

[thestraycat]

Thanks for taking the time to look John.. Appreciate that.

 

Your right it does have 2 NIC's im going to try the other nic now.. But as there both the same NIC i'm assuming the same result tbh.

[John_M]

They are two different chips, according to this: https://www.supermicro.com/products/motherboard/xeon/c202_c204/x9scm-f.cfm

 

4. Intel® 82579LM and 82574L,

    2x Gigabit Ethernet LAN ports

[thestraycat]

Tried the 2nd nic but no joy. Didn't auto negotiate the ip. No VGA cable to work locally on the box until tomorrow but so strange after everything was working any big firmware changes in 6.2.3?

[John_M]

Post your diagnostics zip when you get connected.

 

[thestraycat]

Hmm ok so found a VGA cable, went into GUI mode, noticed that the ethernet adaptors had renamed themselves (eth0 and eth5) refreshed eth0 config hopeing it would resave over the file is saves too... eth5 i gave a static ip so it had one in case i needed to hop over to it. Noticed that my mac addresses for eth0 and eth5 had swapped over in unraid ... switched them back over... rebooted back up for now... Super weird...will keep my eye on it and post in 24hrs. I'd expect an issue in that time.

[JorgeB]

Hmm ok so found a VGA cable, went into GUI mode,

 

Just curious, you are aware your board has IPMI, right?

[thestraycat]

Yup. But the weird thing was that whilst the server was playing up the IPMI was knocked out as well i'm scared to plug it back in tbh now as i removed it whilst troubleshooting - I'm wondering whether the combination of having 2 onboard NIC and 1 ipmi port is a confusing to unraid at times like unraid version upgrades etc.

 

There's no reason tin hell that unraid should be messing with the IPMI port - It felt like it was confused over MAC addresses and was blocking EVERYTHING to the server whilst scanning or trying to resolve.

 

If anyone else has this board i'd love to know your experiences...

[JorgeB]

IPMI port is not detected by unRAID.

[thestraycat]

Very strange. Then i have no idea why it was blocked, tried it multiple times after reboots, hard off the server, punted the switching and router just connected IPMI, then bought the server online. Nothing.

 

Going to try it again in a minute and see if it miraciously seems to work now.

 

Any ideas?

 

Just noticed (whilst i have the server online and working)... When i start a docker in the unraid logs i see:

 

Nov 6 16:50:48 MediaServer kernel: device vethb876690 entered promiscuous mode

Nov 6 16:50:48 MediaServer kernel: eth0: renamed from vethd0005b0

Nov 6 16:50:48 MediaServer kernel: docker0: port 1(vethb876690) entered forwarding state

Nov 6 16:50:48 MediaServer kernel: docker0: port 1(vethb876690) entered forwarding state

Nov 6 16:51:03 MediaServer kernel: docker0: port 1(vethb876690) entered forwarding state

Nov 6 16:51:07 MediaServer kernel: vethd0005b0: renamed from eth0

Nov 6 16:51:07 MediaServer kernel: docker0: port 1(vethb876690) entered disabled state

Nov 6 16:51:07 MediaServer kernel: docker0: port 1(vethb876690) entered disabled state

Nov 6 16:51:07 MediaServer kernel: device vethb876690 left promiscuous mode

Nov 6 16:51:07 MediaServer kernel: docker0: port 1(vethb876690) entered disabled state

 

Is that normal behaviour??

 

[JorgeB]

That's normal.

[John_M]

Yes, that's normal behaviour when you start and stop dockers as virtual NICs are set up and taken down.

 

Which version of unRAID are you running? 6.2 onwards allows you full control of all your NICs (excluding the IPMI port, of course) via the gui. You can force one or other to be eth0 and you can make the one that's reporting as eth5 into eth1.

[thestraycat]

Ah right.. But why the renaming comment

 

"eth0: renamed from beth8ebe344"

 

Working back from all my changes over the last few days and noticed the following RED error in the Linuxserver.io transmission log when stopping it... anyone else have it?

 

 

Nov 6 17:10:13 MediaServer kernel: transmission-da[26536]: segfault at 48 ip 0000556061c5ee19 sp 00002ae9ef40b9e8 error 4 in transmission-daemon[556061c2e000+72000]

 

 

[thestraycat]

Upgraded to Unraid 6.2.4 and it seems to be running ok at the moment... No IPMI still though. Have a funny feeling i'll have to do a full powerdown with router off to reset the BMC chip properly...

 

UPDATE: Yup. Power off of everything bought it back up.