read error on dmesg

[frankz]

Hello everyone I have a cluster with 3 nodes version 8.4 , never had a particular problem . For and today I have a message in the syslog :

[ 4804.474290] vmbr0: port 7(tap112i0) entered disabled state
[ 4804.474507] tap112i0: entered allmulticast mode
[ 4804.476086] vmbr0: port 7(tap112i0) entered blocking state
[ 4804.476090] vmbr0: port 7(tap112i0) entered forwarding state
[ 5484.862257] hrtimer: interrupt took 6088 ns
[ 7598.852714] perf: interrupt took too long (2507 > 2500), lowering kernel.perf_event_max_sample_rate to 79000
[13076.953551] r8169 0000:06:00.0 enp6s0: NETDEV WATCHDOG: CPU: 0: transmit queue 0 timed out 5598 ms
[13076.953571] r8169 0000:06:00.0 enp6s0: ASPM disabled on Tx timeout
[13232.148538] perf: interrupt took too long (3219 > 3133), lowering kernel.perf_event_max_sample_rate to 62000

The cluster system is updated, I think it's a driver problem maybe with the latest kernels.

Can someone give me any information or indications to understand if it is a software problem or other, or disable some option to remedy the error? The cluster runs regularly , migration and more .

 

[MarkusKo]

You could try:
- If everything works just ignore the messages
- Update your nic firmware / bios / uefi
- Disable ASPM in bios (higher power draw?)
- Disable ASPM for your nic (not sure how to)

 

[frankz]

You could try:
- If everything works just ignore the messages
- Update your nic firmware / bios / uefi
- Disable ASPM in bios (higher power draw?)
- Disable ASPM for your nic (not sure how to)

Thank you , but I would like to the origin of problem

 

[MarkusKo]

I'm not an expert but it seems that the r8169 kernel driver tries to set some ASPM mode but the nic does not support it.
Did you upgrade your kernel or change anything?

Here is a vanilla kernel bug report, could be related.