The network falls off on ESXi 6.7 periodically. How to diagnose?

Question

network hardware

Good day, colleagues!
There is a server for a small office, several virtual machines work on it. Hypervisor - free ESXi 6.7.

Configuration

Процессор Intel Xeon E5-2620 v4 LGA 2011-3 20Mb 2.1Ghz
Радиатор SuperMicro SNK-P0048AP4
4 * Память DDR4 Kingston KVR24R17S8/8 8Gb DIMM ECC Reg PC4-19200 CL17 2400MHz
2 * Жесткий диск WD Original SATA-III 2Tb WD2005FBYZ Gold (7200rpm) 128Mb 3.5"
Корпус SuperMicro CSE-732D2-500B
Материнская Плата SuperMicro MBD-X10SRL-F-O Soc-2011 iC612 ATX 8xDDR4 10xSATA3 SATA RAID i210 2хGgbEth Ret
Устройство чтения/записи DVD/CD дисков ASUS DVD±RW+CD/RW DRW-24D5MT/BLK/B/AS black SATA OEM

The problem is this: the network periodically falls off . Healed by reboot.
When this happened for the first time, I immediately went to the office, connected the monitor, the keyboard - there is a picture on the screen, menu items are available. That is, the host did not hang , but neither the host nor the guest OSes are available on the network.
I did not find anything in the logs after (reviewed, everything seems to be).

The first time this happened about a month after the launch, recently more often: once every 1-2 weeks. And it happens on weekends . They tell me about it on Monday morning, or on Saturday, Sunday, if someone went out to work on the weekend. I still don’t see how to compare this pattern with the problem ....

What has been done:
I changed the settings in bios in accordance with the recommendation of arruah here I did not find the IOMMU
item . In accordance with what is indicated here , I included the ASPM parameter . But nothing has changed. There is an option to plug a discrete network card from Intel into the server (on the old server (normal desktop) everything worked with this card for about a year, but, really, there was ESXI 5). But the box is still under warranty, its opening is, as it were, not prohibited, but not desirable, it is sealed.

Answer 1

New firewood on the network did not happen by chance? I had something similar, Windows installed the drivers itself - I looked to see if there are new ones - yes, updated and everything became normal

Answer 2

Motherboard compatibility leaves a lot to be desired.
https://www.supermicro.com/support/resources/OS/C6...
What was wrong with ESXi 5.5?

Answer 3

Hello!
How did you manage to solve the problem?
We have an ESXI 6.7 host, everything worked without failures for about a year, then the network on the host also falls off, and all the adapters at once, no matter what VLAN they are in or what switches are plugged into.
I also noticed a pattern, if by (not using the virtual machine or the host) there is traffic on the network (you download about 10GB 1 file), then the network on ESXI falls, only rebooting the host helps, or pulling / inserting the network wire into the network card.

Answer 4

Problem not solved? I have about the same story only on 6.5. The connection on all interfaces just disappears, it is treated only by reboot.

Answer 5

Friends, and in fact it was possible to solve the problem. Can you share a method?
There was one provider, connected the second. We bought an Intel Original Network Card (EXPI9301CTBLK 893647). They plugged it in, everything started spinning, the connection worked, but after a while the network disappears.
In the "Physical NICs" section, the status of the new interface changes from "1000 Mbps, full duplex" to "Link down".
After restarting the server, the connection is restored, but after a while it disappears again.
in /var/log/vmkernel.log the following entry:

spoiler

2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Insert VLAN Tag'
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing uplink config
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing adapter config
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Strip VLAN Tag'
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing uplink config
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing adapter config
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Xmit Scatter-Gathered Across Multiple Pages'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Offload Checksum for IPv6'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Offload TCP Segmentation for IPv6'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Disabled 'Driver Requires No Packet Scheduling'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Disabled 'Capable To Xmit Scatter-Gathered Data'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Disabled 'Capable To Offload Checksum for IPv4'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Disabled 'Capable To Offload TCP Segmentation for IPv4'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Disabled 'Capable To Insert VLAN Tag'
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing uplink config
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing adapter config
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Disabled 'Capable To Strip VLAN Tag'
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing uplink config
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing adapter config
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Disabled 'Capable To Xmit Scatter-Gathered Across Multiple Pages'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Disabled 'Capable To Offload Checksum for IPv6'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Disabled 'Capable To Offload TCP Segmentation for IPv6'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Driver Requires No Packet Scheduling'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Xmit Scatter-Gathered Data'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Offload Checksum for IPv4'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Offload TCP Segmentation for IPv4'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Insert VLAN Tag'
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing uplink config
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing adapter config
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Strip VLAN Tag'
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing uplink config
2020-11-14T10:02:02.152Z cpu1:2097220)DEBUG (ne1000): writing adapter config
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Xmit Scatter-Gathered Across Multiple Pages'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Offload Checksum for IPv6'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Enabled 'Capable To Offload TCP Segmentation for IPv6'
2020-11-14T10:02:02.152Z cpu1:2097220)INFO (ne1000): vmnic2: Disabled 'Driver Requires No Packet Scheduling'
2020-11-14T10:02:02.152Z cpu2:2097296)CpuSched: 699: user latency of 2113612 vmnic2-0-tx 0 changed by 2097296 NetSchedHelper -6
2020-11-14T10:02:02.152Z cpu0:2113612)NetSched: 654: vmnic2-0-tx: worldID = 2113612 exits
2020-11-14T10:02:02.152Z cpu2:2097296)CpuSched: 699: user latency of 2113613 vmnic2-0-tx 0 changed by 2097296 NetSchedHelper -6
2020-11-14T10:02:03.604Z cpu0:2097609)DEBUG (ne1000): vmnic2: retry to wait for link up
2020-11-14T10:02:05.604Z cpu0:2097609)DEBUG (ne1000): vmnic2: retry to wait for link up
2020-11-14T10:02:07.604Z cpu2:2097609)DEBUG (ne1000): vmnic2: retry to wait for link up
2020-11-14T10:02:21.605Z cpu2:2097609)INFO (ne1000): vmnic2: Link is Down
2020-11-14T10:02:21.605Z cpu2:2097609)DEBUG (ne1000): Reporting uplink 0x4304ad48e950 status
2020-11-14T10:07:49.149Z cpu1:2097693)DVFilter: 5963: Checking disconnected filters for timeouts
2020-11-14T10:17:49.147Z cpu6:2097693)DVFilter: 5963: Checking disconnected filters for timeouts

Answer 6

changed the poppy address of the virtual machine, the first 6 characters can be changed and everything started up.