HP DL 380 Gen 9 issues on 5.4.73 & 5.4.78 kernel

[sillyvainplex]

Sorry that it took so long to reply. I also did a Proxmox 6.3 upgrade on my HP BL460c G6 and it also made it the Proxmox website unreachable.

So I guess it affects G6's and G8's and I wonder if it will do the same on my G9's

 

[NTS]

Hy
We are planning to upgrade our Proxmox Cluster, based on HP DL380 G9 Servers. But as i can read, its a bad idea.
ATM the upgrade process would install the Kernel 5.4.78-2.
When is the tested Kernel pve_5.4.78-1~b1_amd64 available?
Thank you very much!

Cheers A

 

[t.lamprecht]

ATM the upgrade process would install the Kernel 5.4.78-2.

That is the fixed kernel, as mentioned in this thread already.

When is the tested Kernel pve_5.4.78-1~b1_amd64 available?

That was just a test build, 5.4.78-2 is newer and includes those fixes and was released to all repositories since start of December 2020.

 

[NTS]

That is the fixed kernel, as mentioned in this thread already.


That was just a test build, 5.4.78-2 is newer and includes those fixes and was released to all repositories since start of December 2020.

Thank you very much for the update! In this case, we can run the upgrade to 6.3.

 

[sillyvainplex]

Hey I wanted to say that uploading the custom kernel fixed the issue. The BL460c Gen 6 blade servers are not affected which the installation was fine without any hiccups. The issue is with the BL460c Gen 8 and Gen 9's which it's all good now thanks to you.

 

[NTS]

Ah, i like to say the same - for next visitors:
HP DL380 G9 works flawless with 5.4.78-2

 

[encore]

I can confirm it works with kernel 5.4.78-2 without any problems on HP BL460c g9, but not with newer kernel 5.4.140-1-pve.
Why does the newest kernel does not include the fixes @t.lamprecht ? I don't want to stuck on 5.4.78-2 forever, especially because it comes with ceph-common 12 but I need ceph-common 15.

 

[t.lamprecht]

I can confirm it works with kernel 5.4.78-2 without any problems on HP BL460c g9, but not with newer kernel 5.4.140-1-pve.
Why does the newest kernel does not include the fixes @t.lamprecht ?

The problematic patch is still reverted with 5.4.140... I'd guess you run into another issue, what is not working, what errors do you get?
Quite some 5.4 based kernels have been released in-between to, so maybe try those to narrow down where the new issue that you observe has been introduced.

I don't want to stuck on 5.4.78-2 forever, especially because it comes with ceph-common 12 but I need ceph-common 15.

What? Ceph and the kernel are rather independent, you can update the whole system but boot an older kernel...

 

[encore]

symptoms are quite the same as described in the thread here. If I boot with 5.4.78-2 network is working fine. I see an eno49 and eno50 network device. When I boot with 5.4.140 no network devices can be found anymore:

Attached you find the DMESG, journal and syslog outputs.

Btw, I can also confirm it only affects bl460c g9, our bl460c g8 systems are working fine with 5.4.140.

 

[encore]

I compared the relevant part of journalctl with the working 5.4.78-2 kernel and the non-working 5.4.140 kernel:

not working 5.4.140:

Sep 26 19:29:32 captive072-48000-bl09 kernel: bnx2x: probe of 0000:06:00.0 failed with error -22
Sep 26 19:29:32 captive072-48000-bl09 kernel: bnx2x 0000:06:00.1: msix capability found
Sep 26 19:29:32 captive072-48000-bl09 kernel: bnx2x 0000:06:00.1: part number 0-0-0-0
Sep 26 19:29:32 captive072-48000-bl09 kernel: bnx2x: probe of 0000:06:00.1 failed with error -22

working 5.4.78-2:

Sep 26 19:34:59 captive072-48000-bl09 kernel: bnx2x 0000:06:00.0: 63.008 Gb/s available PCIe bandwidth (8 GT/s x8 link)
Sep 26 19:34:59 captive072-48000-bl09 kernel: bnx2x 0000:06:00.1: msix capability found
Sep 26 19:34:59 captive072-48000-bl09 kernel: bnx2x 0000:06:00.1: part number 0-0-0-0
Sep 26 19:34:59 captive072-48000-bl09 kernel: bnx2x 0000:06:00.1: 63.008 Gb/s available PCIe bandwidth (8 GT/s x8 link)
Sep 26 19:34:59 captive072-48000-bl09 kernel: bnx2x 0000:06:00.1 eno50: renamed from eth1

any ideas @t.lamprecht ?

 

[sillyvainplex]

I don't have any issues now. everything is working great

 

[encore]

i have replaced the network adapter 536 FLB (broadcom based) with a 560 flb (intel based). Since then it runs with the latest proxmox kernel.
536FLB runs only with the old kernel 5.4.78-2