Latency Test OK, dennoch Fehlermeldung
- tommylight
- Online
- Moderator
Less
More
- Posts: 18726
- Thank you received: 6296
05 Aug 2024 13:29 #306994
by tommylight
Replied by tommylight on topic Latency Test OK, dennoch Fehlermeldung
What do you have now in etc/network/interfaces ?
The following user(s) said Thank You: Muftijaja
Please Log in or Create an account to join the conversation.
05 Aug 2024 13:38 #306995
by Muftijaja
Replied by Muftijaja on topic Latency Test OK, dennoch Fehlermeldung
/etc/network/interfaces has just
auto lo
iface lo inet loopback
everything else I did outcomment for the moment, but I need to have the better performance for the intel card
auto lo
iface lo inet loopback
everything else I did outcomment for the moment, but I need to have the better performance for the intel card
Please Log in or Create an account to join the conversation.
- tommylight
- Online
- Moderator
Less
More
- Posts: 18726
- Thank you received: 6296
05 Aug 2024 13:57 #306996
by tommylight
Replied by tommylight on topic Latency Test OK, dennoch Fehlermeldung
OK, add as PCW said for you Intel NIC, save and reboot, see if it works with Mesa, double check with ip a and ping.
The following user(s) said Thank You: Muftijaja
Please Log in or Create an account to join the conversation.
05 Aug 2024 14:02 #306997
by Muftijaja
Replied by Muftijaja on topic Latency Test OK, dennoch Fehlermeldung
Ok, what adress to use - 10.10.10.2 or 10.10.10.1?
Please Log in or Create an account to join the conversation.
05 Aug 2024 14:20 #307000
by Muftijaja
Replied by Muftijaja on topic Latency Test OK, dennoch Fehlermeldung
so, I tried both adresses, with both settings in etc/network/interfaces I have no networks afterwards.
Hiernach geht die Verbindung gar nicht mehr:
auto eno1
iface eno1 inet static
address 10.10.10.2 (same with 10.10.10.1)
hardware-irq-coalesce-rx-usecs 0
Hiernach geht die Verbindung gar nicht mehr:
auto eno1
iface eno1 inet static
address 10.10.10.2 (same with 10.10.10.1)
hardware-irq-coalesce-rx-usecs 0
Please Log in or Create an account to join the conversation.
05 Aug 2024 20:57 #307031
by PCW
Replied by PCW on topic Latency Test OK, dennoch Fehlermeldung
So I tried with a PC with Debian 12 and LinuxCNC 2.9.2 and
everything works as expected. I did setup the real time interface
in the network manager and just added the interfaces file settings
for IRQ coalescing.
OS version
Linux EliteDesk800G3 6.1.0-18-rt-amd64 #1 SMP PREEMPT_RT Debian 6.1.76-1 (2024-02-01) x86_64 GNU/Linux
Debian release:
interfaces:
ip a result:
everything works as expected. I did setup the real time interface
in the network manager and just added the interfaces file settings
for IRQ coalescing.
OS version
Linux EliteDesk800G3 6.1.0-18-rt-amd64 #1 SMP PREEMPT_RT Debian 6.1.76-1 (2024-02-01) x86_64 GNU/Linux
Debian release:
interfaces:
ip a result:
Attachments:
The following user(s) said Thank You: Muftijaja
Please Log in or Create an account to join the conversation.
05 Aug 2024 23:22 #307044
by Muftijaja
Replied by Muftijaja on topic Latency Test OK, dennoch Fehlermeldung
THanks for your simulation check with a new installation, PCW!
OK, it seems that my network configuration is magically faulty and can not be fixed to work with the lines in etc/network/interfaces, to reduce the ping times to my Mesa card in the normal way.
But, I don't want to install everything including Debian and LCNC completely new.
I think, I have to bite in the sour apple and fix it with the line
sudo ethtool -C eno1 rx-usecs 0 before each and every start of my LCNC Machine. I know that this works.
That wouldn't be comfortable but it works. And I can go further with my project.
My last step was to delete everything than the loop connection 127.0.0.1 from the /interfaces file,
configuring the Intel card with the network manager as before, and using the RTL8111 connection with Auto Internet.
The ip a with this connection is as follows:
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host noprefixroute
valid_lft forever preferred_lft forever
2: enp2s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
link/ether 00:0a:cd:00:00:56 brd ff:ff:ff:ff:ff:ff
inet 192.168.178.55/24 brd 192.168.178.255 scope global dynamic noprefixroute enp2s0
valid_lft 831324sec preferred_lft 831324sec
inet6 2003:ce:5f08:8600:d50e:a6ab:524:e6ae/64 scope global dynamic noprefixroute
valid_lft 6923sec preferred_lft 1390sec
inet6 fe80::8e97:e3df:28fe:e8cf/64 scope link noprefixroute
valid_lft forever preferred_lft forever
3: eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
link/ether 00:23:24:5b:b4:ac brd ff:ff:ff:ff:ff:ff
altname enp0s25
inet 10.10.10.2/24 brd 10.10.10.255 scope global noprefixroute eno1
valid_lft forever preferred_lft forever
What do you think? Any other solution?
Thanks again for your thougts and hints!
Hanno/arno
OK, it seems that my network configuration is magically faulty and can not be fixed to work with the lines in etc/network/interfaces, to reduce the ping times to my Mesa card in the normal way.
But, I don't want to install everything including Debian and LCNC completely new.
I think, I have to bite in the sour apple and fix it with the line
sudo ethtool -C eno1 rx-usecs 0 before each and every start of my LCNC Machine. I know that this works.
That wouldn't be comfortable but it works. And I can go further with my project.
My last step was to delete everything than the loop connection 127.0.0.1 from the /interfaces file,
configuring the Intel card with the network manager as before, and using the RTL8111 connection with Auto Internet.
The ip a with this connection is as follows:
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host noprefixroute
valid_lft forever preferred_lft forever
2: enp2s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
link/ether 00:0a:cd:00:00:56 brd ff:ff:ff:ff:ff:ff
inet 192.168.178.55/24 brd 192.168.178.255 scope global dynamic noprefixroute enp2s0
valid_lft 831324sec preferred_lft 831324sec
inet6 2003:ce:5f08:8600:d50e:a6ab:524:e6ae/64 scope global dynamic noprefixroute
valid_lft 6923sec preferred_lft 1390sec
inet6 fe80::8e97:e3df:28fe:e8cf/64 scope link noprefixroute
valid_lft forever preferred_lft forever
3: eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
link/ether 00:23:24:5b:b4:ac brd ff:ff:ff:ff:ff:ff
altname enp0s25
inet 10.10.10.2/24 brd 10.10.10.255 scope global noprefixroute eno1
valid_lft forever preferred_lft forever
What do you think? Any other solution?
Thanks again for your thougts and hints!
Hanno/arno
Please Log in or Create an account to join the conversation.
06 Aug 2024 00:36 #307048
by PCW
Replied by PCW on topic Latency Test OK, dennoch Fehlermeldung
Really odd, I did verify that the IRQ coalescing setting worked.
The only thing I can think of of is that there's a bad character etc
in the interfaces file (you might try using my interfaces file)
This is possible since any errors parsing the file are well hidden
Another possibility is some strange interaction between the
network manager settings and the interfaces file. This could
be tested by deleting the network manager setup for the realtime
connection. I think something may be broken in this interaction
with newer Debian systems as older ones would not allow you to
change settings for an interface that was setup in the interfaces file
(it would be marked as "Unmanaged")
The only thing I can think of of is that there's a bad character etc
in the interfaces file (you might try using my interfaces file)
This is possible since any errors parsing the file are well hidden
Another possibility is some strange interaction between the
network manager settings and the interfaces file. This could
be tested by deleting the network manager setup for the realtime
connection. I think something may be broken in this interaction
with newer Debian systems as older ones would not allow you to
change settings for an interface that was setup in the interfaces file
(it would be marked as "Unmanaged")
The following user(s) said Thank You: Muftijaja
Please Log in or Create an account to join the conversation.
06 Aug 2024 10:28 - 06 Aug 2024 10:37 #307067
by Muftijaja
Replied by Muftijaja on topic Latency Test OK, dennoch Fehlermeldung
PCW, thanks for your ideas!
Well, I tried copy&Paste with your interfaces file, but no change.
After that, I believe that your idea of smth broken in the interaction of interfaces file and network manager is true. My last try today was to delete (again) any connection in the network manager and restart with the interfaces file. The result was as before - no eno1 connection possible. that's really crazy.
On the other hand, deleting the entries in interface file and managing the eno1 connection I got the connection to the mesa card but bad latency (not so bad but not sufficient). So, I am facing the decision to make a complete new Debian and LCNC installation or living with that.
Q: Is it possible to put the lines for ethtool in a textfile and start it from the desktop (like a batch in DOS/Windows) ? That would make it a little more comfortable. At the moment I don't want to take the lot of trouble with a full new installation.
BTW one more side effect - If I switch the mesa connection with the ethtool to -C eno1 rx-usecs 0 the auto interne connection does not work anymore. Connection is shown, but no function.
Q: Ok, I don't want to surf the internet while cnc'ing, but how can I stop/reset the ethtool enty?
Thanks for your answers!
Well, I tried copy&Paste with your interfaces file, but no change.
After that, I believe that your idea of smth broken in the interaction of interfaces file and network manager is true. My last try today was to delete (again) any connection in the network manager and restart with the interfaces file. The result was as before - no eno1 connection possible. that's really crazy.
On the other hand, deleting the entries in interface file and managing the eno1 connection I got the connection to the mesa card but bad latency (not so bad but not sufficient). So, I am facing the decision to make a complete new Debian and LCNC installation or living with that.
Q: Is it possible to put the lines for ethtool in a textfile and start it from the desktop (like a batch in DOS/Windows) ? That would make it a little more comfortable. At the moment I don't want to take the lot of trouble with a full new installation.
BTW one more side effect - If I switch the mesa connection with the ethtool to -C eno1 rx-usecs 0 the auto interne connection does not work anymore. Connection is shown, but no function.
Q: Ok, I don't want to surf the internet while cnc'ing, but how can I stop/reset the ethtool enty?
Thanks for your answers!
Last edit: 06 Aug 2024 10:37 by Muftijaja.
Please Log in or Create an account to join the conversation.
06 Aug 2024 14:41 - 06 Aug 2024 14:42 #307093
by PCW
Replied by PCW on topic Latency Test OK, dennoch Fehlermeldung
So you are saying that changing eno1 coalesce settings affects the other interface?
That's extremely weird
Is it possible both interfaces are sharing an interrupt?
(lshw should show this)
That's extremely weird
Is it possible both interfaces are sharing an interrupt?
(lshw should show this)
Last edit: 06 Aug 2024 14:42 by PCW.
Please Log in or Create an account to join the conversation.
Time to create page: 0.096 seconds