Skip to main content ↑ ↓
Topic: etho keeps dropping and can't restart (Read 4036 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Re: etho keeps dropping and can't restart

Reply #15 –
Quote from: Artist – on
This is the 3rd report (the 1st 2 not in this forum) I noticed in the past 10 days about this issue.
It might very well be related to a kernel change.

This could very well be related to https://bugzilla.kernel.org/show_bug.cgi?id=118721

Comments 9 and 10 show workarounds:

Quote
Disabling TSO seems to have fixed the problem for me. (I needed to set it after a fresh boot, *before* the interface starts bailing out continually.)

Quote
Also hit this issue - might be helpful to others, reloading the module with the parameter Node=0 (The NUMA node my NIC is on - modprobe e1000e Node=0) appears to have worked around the issue.

I suggest to try these, also to clarify if it indeed is the same problem as reported here.

artist


How would I go about doing this?

Re: etho keeps dropping and can't restart

Reply #18 –
any news on a real fix for this.  This is an old kernel level problem

Re: etho keeps dropping and can't restart

Reply #19 –
On the only system we had such a problem with we found out that the network service was started before the eth0 interface was ready, After having added a wait loop to work around this the problem did not occur again afaik.

artist

Re: etho keeps dropping and can't restart

Reply #20 –
Quote from: artoo – on
Please also test linux-6.13 in gremlins.
My first test with it seems to have the problem fixed.

Quote from: artoo – on
Please also test linux-6.13 in gremlins.
My first test with it seems to have the problem fixed.



I just ran an update on the webserver and it was very hard... the ethernet just crashed .  It was so fustrating I just pulled the Damn plug on it and did a hard reboot.  I then ran

sudo ethtool -K eth1 gso off gro off tso off

It is up now but I don't know how stable.

This update mariabd switched from mysqld to mariad - and that failed to go smoothly.  It is up now.

I hate ghost hardware problems.  My guess is that the modules are all there because I made an custom module alteration for the ethernet card - years ago.  it is looking for the source code still, even though the current module is upto date.