Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

after updated to 24.5.3 can't connect. #963

Open
eragon4k opened this issue May 12, 2024 · 65 comments
Open

after updated to 24.5.3 can't connect. #963

eragon4k opened this issue May 12, 2024 · 65 comments

Comments

@eragon4k
Copy link

eragon4k commented May 12, 2024

It shows and detected e1000e (intel I219-LM) but cant connect.

请填写以下信息.
Please fill in the following information.

Install ENV: (You can find it in the boot interface.)

  • DMI: Thinkstation p520
  • CPU: Xeon W-2135
  • NIC: intel I219-LM

RR version: (You can find it in the update menu.)

  • RR: 24.5.3
  • addons:
  • modules:
  • lkms:

DSM:

  • model: DS3622xs+
  • version: 7.2.1-69057 update 5

Issue: can't connect after updating from 24.5.1 to 24.5.3 and 24.5.4 with (Priority use of official drivers: false or Priority use of official drivers: ture)

logs:

(请先看一下#173#175、#226的内容)
(Plz review the content of #173, #175, #226 first)
...

... 如果你提供不了详细信息,那就等有缘人吧!
... If you can't provide detailed information, then wait for someone who is destined!

@wjz304
Copy link
Contributor

wjz304 commented May 13, 2024

check other model.
Or enable this options.
image

@eragon4k
Copy link
Author

eragon4k commented May 13, 2024

used RR Manager 2.0.19 from 24.5.1 to 24.5.3.

24.5.1 was working and stable. (checked and it was working with "Priority use of official drivers: false")

I'll give it a try "Priority use of official drivers: true" method.

@eragon4k
Copy link
Author

Looks like it's not the driver issue.
I believe there is conflict when you setup static ip with in the DSM.

When the loader give and detects lan and it give initial IP address to connect but when you have static ip already set up inside DSM. It not responding from static ip and the loaders initial ip address.

And when it boots, I ping the loader's initial IP address, and it responds, and then it stops.

Now I'm going back to 24.5.1, removing the static IP from the DSM, and trying to update.

@wjz304
Copy link
Contributor

wjz304 commented May 14, 2024

The e1000e simulated by pve will continue to go up/down using the driver I compiled. Just use the official driver.
Of course, there are many models of e1000e, and it may not be suitable for all.
Just that the e1000e driver has not been updated recently.

@eragon4k
Copy link
Author

eragon4k commented May 14, 2024

i did tried "Priority use of official drivers: true" with 24.5.4 but the result was the same thing.

I'll try to rebuild the loader from scratch.

@eragon4k
Copy link
Author

So I did a few tests.

  1. I freshly built the loader and the "Priority use of official drivers: true" method with 24.5.4.
    Out comes with no success.

  2. With version 24.5.1 removed DSM Static ip and updated to 24.5.4 with "Priority use of official drivers: true" and out come was no success.

For now going back to 24.5.1.

@wjz304
Copy link
Contributor

wjz304 commented May 15, 2024

取一下 pid vid

@eragon4k
Copy link
Author

for usb drive?

@wjz304
Copy link
Contributor

wjz304 commented May 15, 2024

intel I219-LM

@wjz304
Copy link
Contributor

wjz304 commented May 15, 2024

#173 (comment)

@wjz304
Copy link
Contributor

wjz304 commented May 15, 2024

find [0200]

@eragon4k
Copy link
Author

Can't find it.

The log is scrolling far too quickly.

There may be more logs on the top part of the screen.

@wjz304
Copy link
Contributor

wjz304 commented May 15, 2024

lspci -nn | grep 0200

@eragon4k
Copy link
Author

eragon4k commented May 15, 2024

I219-LM [8086:15b7]

Thank you

@snailium
Copy link

My loader lost model/architecture information after upgrade to 24.5.4. I had to rebuild the loader from scratch.

After that, DSM seems corrupted. It gave me the first setup wizard. I had to setup everything as new, and recover the configuration from Synology account.

There are too much hassle upgrade cross 24.5.2 version, which seems heavily restructuring in the loader.

@wjz304
Copy link
Contributor

wjz304 commented May 16, 2024

升级到 24.5.4 后,我的加载器丢失了模型/架构信息。我不得不从头开始重建装载机。

在那之后,DSM似乎已损坏。它给了我第一个设置向导。我必须将所有内容都设置为新的,并从 Synology 帐户中恢复配置。

跨 24.5.2 版本升级有太多麻烦,这似乎在加载器中进行了大量重组。

My loader lost model/architecture information after upgrade to 24.5.4. I had to rebuild the loader from scratch.

After that, DSM seems corrupted. It gave me the first setup wizard. I had to setup everything as new, and recover the configuration from Synology account.

There are too much hassle upgrade cross 24.5.2 version, which seems heavily restructuring in the loader.

Yes, the logic adjustment in 5.2 is significant, but it does not involve driver adjustment,
At that time, I was hesitant about whether to revise it to restrict updates from being allowed,

@wjz304
Copy link
Contributor

wjz304 commented May 18, 2024

I219-LM [8086:15b7]

Thank you

还好,官方的驱动也支持这个 pid&vid

@eragon4k
Copy link
Author

I219-LM [8086:15b7]
Thank you

Fortunately, the official driver also supports this pid&vid

So, my understanding is the driver was removed after 24.5.1. Right?

@wjz304
Copy link
Contributor

wjz304 commented May 18, 2024

I219-LM [8086:15b7]
Thank you

Fortunately, the official driver also supports this pid&vid

So, my understanding is the driver was removed after 24.5.1. Right?

no, I don’t know why at the moment. I’ll try updating the driver version.

@eragon4k
Copy link
Author

I219-LM [8086:15b7]
Thank you

Fortunately, the official driver also supports this pid&vid

So, my understanding is the driver was removed after 24.5.1. Right?

no, I don’t know why at the moment. I’ll try updating the driver version.

I did try with Priority use of official drivers: true as you mentioned before

@lyp49472060
Copy link

24.4.6升级到24.5.4一样找不到

@wjz304
Copy link
Contributor

wjz304 commented May 18, 2024

try add e1000e.KumeranLockLoss=1 to cmdline
image

@eragon4k
Copy link
Author

try add e1000e.KumeranLockLoss=1 to cmdline image

Is this reply for me or @lyp49472060?

@wjz304
Copy link
Contributor

wjz304 commented May 18, 2024

you

try add e1000e.KumeranLockLoss=1 to cmdline image

Is this reply for me or @lyp49472060?

@eragon4k
Copy link
Author

you

try add e1000e.KumeranLockLoss=1 to cmdline image

Is this reply for me or @lyp49472060?

Understood. I'll try this later today and update you.

Thank you

@eragon4k
Copy link
Author

eragon4k commented May 18, 2024

you

try add e1000e.KumeranLockLoss=1 to cmdline image

Is this reply for me or @lyp49472060?

so i tried with "e1000e.KumeranLockLoss=1". the result was the same and I just removed the ethernet cable and re-plugged it and it was getting a ping again.

therefore i removed "e1000e.KumeranLockLoss=1" and test it out to see if I could get a ping just by removing the ethernet cable and re-plugging it. Surprisingly I was getting a ping.

But after the recovery process it’s not responding even after unplugging and relugging the ethernet cable.

@wjz304
Copy link
Contributor

wjz304 commented May 19, 2024

How many NIC are there in total?

@eragon4k
Copy link
Author

How many NIC are there in total?

Just one.
I am using a white label SN but it got banned after upgrade to 5.4.

@wjz304
Copy link
Contributor

wjz304 commented May 19, 2024

e1000e has another parameter e1000e.SmartPowerDownEnable=1 , You can also try it

@wjz304
Copy link
Contributor

wjz304 commented May 19, 2024

I use PVE to simulate e1000e, with a total of 5 simulated network cards. After adding e1000e.KumeranLockLoss=1 and netifsort addon, DHCP to IP works normally

@eragon4k
Copy link
Author

try RR shell sed -i 's|/etc/init.d/S41dhcpcd|#/etc/init.d/S41dhcpcd|g' /opt/rr/boot.sh /opt/rr/init.sh

I had an expi9301ctblk and just added 9301 NIC and it works for now with 5.6.
Do you know if the intel 1226v or rtl8125b is supported?

What is your recommendation for a 2.5g nic?

@wjz304
Copy link
Contributor

wjz304 commented May 30, 2024

try RR shell sed -i 's|/etc/init.d/S41dhcpcd|#/etc/init.d/S41dhcpcd|g' /opt/rr/boot.sh /opt/rr/init.sh

I had an expi9301ctblk and just added 9301 NIC and it works for now with 5.6. Do you know if the intel 1226v or rtl8125b is supported?

What is your recommendation for a 2.5g nic?

i226 OK, (No one is currently not available),
R8125* OK, (someone feeds that certain suffixes are unavailable, but no details are provided).

@wjz304
Copy link
Contributor

wjz304 commented May 31, 2024

@eragon4k
Copy link
Author

eragon4k commented May 31, 2024

@eragon4k e1000e try https://github.com/RROrg/rr/releases/tag/24.6.0

It didn't work.
But one wierd behavior.

When it boots, it's not responding from the e1000e, so I plug it into expi9301 and it works, and then while it's turned on, I plug it back into the e1000e and it works.

so it's just not responding when it boots.

Package center package icons are missing with 24.6.0

@nillebor
Copy link

nillebor commented Jun 3, 2024

i'm still with the old version. also the current release does not work

My device: HP Elitedesk G4 800 with Intel I219-LM LAN.
The error occurs only with DSM boot, reconnecting also does not work anything. Everything works as it should in the loader. Of course, clean installation.

Please help and fix it

@wjz304
Copy link
Contributor

wjz304 commented Jun 4, 2024

e1000e.zip
image

@wjz304
Copy link
Contributor

wjz304 commented Jun 4, 2024

To be honest, I don't know why. Since 24.5.1, I haven't modified the driver-related functions. Some issues related to DHCP have also been rolled back in 24.6.0. This problem is a bit difficult to solve.
I don't have a physical e1000e device. I use PVE virtual e1000e. There are no problems in various tests.
I modified the driver again and tested it.

e1000e.zip image

@nillebor
Copy link

nillebor commented Jun 4, 2024

Unfortunately, I have not had any success. I tried 2 DS (gemenilake & broadwellnk) in both versions. The connection works up to the loader. After that, a connection (ping) is no longer possible. Even a reboot does not help.

@nillebor
Copy link

nillebor commented Jun 4, 2024

Static IP in loader does not work.

@wjz304
Copy link
Contributor

wjz304 commented Jun 4, 2024

dmesg | grep e1000e
lsmod | grep e1000e

@snailium
Copy link

snailium commented Jun 4, 2024

Just some idea. Is it related to the SN/MAC settings? How is the MAC used and how does DSM use the supplied MAC value?

Maybe try to set mac1 and mac2 to be the real MAC of the NIC (instead of 00:11:32 ones) and see if that works.

@wjz304
Copy link
Contributor

wjz304 commented Jun 4, 2024

Unfortunately, I have not had any success. I tried 2 DS (gemenilake & broadwellnk) in both versions. The connection works up to the loader. After that, a connection (ping) is no longer possible. Even a reboot does not help.

get pid vid (Used to confirm whether it only occurs in a certain model)

@nillebor
Copy link

nillebor commented Jun 5, 2024

dmesg | grep e1000e:

[   13.660450] e1000e: Intel(R) PRO/1000 Network Driver
[   13.660451] e1000e: Copyright(c) 1999 - 2015 Intel Corporation.
[   13.661506] e1000e 0000:00:1f.6: Interrupt Throttling Rate (ints/sec) set to dynamic conservative mode
[   13.810673] e1000e 0000:00:1f.6 0000:00:1f.6 (uninitialized): registered PHC clock
[   13.876500] e1000e 0000:00:1f.6 eth0: (PCI Express:2.5GT/s:Width x1) xx:xx:xx:xx:xx:xx
[   13.876503] e1000e 0000:00:1f.6 eth0: Intel(R) PRO/1000 Network Connection
[   13.876644] e1000e 0000:00:1f.6 eth0: MAC: 13, PHY: 12, PBA No: FFFFFF-0FF
[   18.547463] e1000e 0000:00:1f.6 eth0: NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None

Duplicate of #The correct Mac address is displayed, I just anonymized it.

lsmod | grep e1000e:
e1000e 282624 0

the same result with true and false option (p):
grafik

@nillebor
Copy link

nillebor commented Jun 5, 2024

I select the DS, DSM and add-ons. Then I build the loader and start it. Under 24.4.8 it works without problems. But I need the current modules because of NVMe. An update of the modules is not possible. The network connection is disconnected. Also in the router, the device is no longer displayed as connected.

grafik

@snailium
Copy link

snailium commented Jun 5, 2024

I select the DS, DSM and add-ons. Then I build the loader and start it. Under 24.4.8 it works without problems. But I need the current modules because of NVMe. An update of the modules is not possible. The network connection is disconnected. Also in the router, the device is no longer displayed as connected.

grafik

What is skip_vendor_mac_interface?

@nillebor
Copy link

nillebor commented Jun 5, 2024

The last working version for me is the 24.5.1. (prerelease not tested).

@snailium,
the setting is default. i have also changed nothing about the model, version and NVMe add-on. the same procedure as in the other versions as well. I just need the NVMe fix.

@snailium
Copy link

snailium commented Jun 5, 2024

The last working version for me is the 24.5.1. (prerelease not tested).

@snailium,
the setting is default. i have also changed nothing about the model, version and NVMe add-on. the same procedure as in the other versions as well. I just need the NVMe fix.

I'm asking because I built my loader from scratch and I don't have that skip vendor parameter. It seems it is network related. Maybe try to get rid of that parameter and see if any improvements?

@nillebor
Copy link

nillebor commented Jun 5, 2024

this option is under cmdline. There is nothing to be found under cmdline, modules and addons.

grafik

@snailium
Copy link

snailium commented Jun 6, 2024

it seems the skip vendor parameter is added in 24.6.0. I don't see this option in older loader.

@nillebor
Copy link

nillebor commented Jun 8, 2024

After the new update 24.6.1, the connection does not work. :(
How can I help to fix the error?

@wjz304
Copy link
Contributor

wjz304 commented Jun 8, 2024

image
Change this option to true and try

@lyp49472060
Copy link

lyp49472060 commented Jun 8, 2024

24.6.1更新后,我的DSM依旧失联。24.4.6版本正常
网卡是intel 82599 10g,固定ip地址,ds920+,

@wjz304
Copy link
Contributor

wjz304 commented Jun 8, 2024

如果方便就远程(anydesk,todesk,向日葵)吧,QQ: 304403268,TG: https://t.me/wjz304

@lyp49472060
Copy link

24.6.1更新后,我的DSM依旧失联。24.4.6版本正常 网卡是intel 82599 10g,固定ip地址,ds920+,

ESXi-8.0U2c 确认和下图一致,
屏幕截图 2024-06-08 185319
336830557-f3a4b212-f087-45b5-bdf5-192c36f90452

删除vmw_pvscsi,正常启动

@nillebor
Copy link

nillebor commented Jun 8, 2024

@wjz304,

Thanks, the booting and connecting works. :)
Can you just briefly say what the problem is and why it doesn't work with the default values?

@wjz304
Copy link
Contributor

wjz304 commented Jun 8, 2024

@wjz304,

Thanks, the booting and connecting works. :) Can you just briefly say what the problem is and why it doesn't work with the default values?

I just guess that it is related to the power management or memory of the network card.
Direct boot will skip the RR system, and the network card will be started under DSM
Indirect boot means that the network card is started under RR, and then kexec to the DSM system to start again. This process may cause power/memory abnormalities of the network card

@wjz304
Copy link
Contributor

wjz304 commented Jun 8, 2024

We still need to investigate the details, but it is not easy to investigate the existing logs for problems like this. I don't have the relevant hardware, so this is difficult.

@nillebor
Copy link

nillebor commented Jun 8, 2024

@wjz304,
This process may cause power/memory abnormalities of the network card

wouldn't it be better to make direct boot as standard?

@wjz304
Copy link
Contributor

wjz304 commented Jun 8, 2024

@wjz304,
This process may cause power/memory abnormalities of the network card

wouldn't it be better to make direct boot as standard?

There is no good or bad difference, the non-direct startup UI just displays more information

@eragon4k
Copy link
Author

eragon4k commented Jun 9, 2024

image Change this option to true and try

I confirm that this method works.
Wish it could display more information, but this is better than not working.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

6 participants