HPE 1920S switch temporary doesn't replies to ICMP (ping) requests

[eldxmgw]

HPE 1920S switch temporary doesn't replies to ICMP (ping) requests if I'm powering on more than two hosts. The hosts are physical ones. Ongoing ping requests from a client shows for example ping timeouts on host 2 while host 3 replies to pings. Then after some seconds (not always the same) host 2 replies while host 3 shows ping timeouts vice versa. Doing so with just two hosts, no problem at all. Powering more than 3 hosts its going even worse. All hosts are in one VLAN, the switch has the latest firmware and showed this scenario already with older ones. I also disabled all green and security features and also checked the switching port configuration of each port connected to every host. This makes me mad because the vCenter looses the connection to the hosts. The problem only happens with the hosts, but no other device so far connected to this 52 port switch. This issue makes no sense to me and i have no idea how to fix this. Somebody has an idea? I found some threads but with no solutions: https://community.hpe.com/.../HP-1920.../td-p/6873613

https://community.hpe.com/.../HPE-1920.../td-p/6973286

[akg7]

Hello @eldxmgw ,

Can you check the spanning tree status?

Also check same host connecting to other ports.

Thanks!

[eldxmgw]

Hello and thanks for your reply.

Switching -> Spanning Tree -> Spanning Tree admin mode is disabled.

The cluster consists of 8 hosts, where every host is connected with 3x RJ-45 wires to this switch. Two data paths in one VLAN, one for IRMC management in another VLAN. So, 15 ports are covered from 8 hosts, and the behavior is the same. I'd say this answers if this happens on other ports.

[akg7]

Hello @eldxmgw,

Is it possible to check with 1-to-1 connection with switch and host instead of 3XRJ45?

Thanks!

[eldxmgw]

Sure, did so already.

Right now i powered on 3 hosts. One connection per host to the switch. Other RJ-45 cables are disconnected. Still the same. I attached a screenshot where you can see that when .11 fails .12 is available vice versa. In this example with 3 powered hosts only .10 is the only stable option. As said before, as more hosts are powered on it wents even more worse. -> https://www.dropbox.com/s/vi02cu68vqw8atf/screenshot.png?dl=0

You mentioned the spanning tree protocol... should something like MSTP be enabled?

[akg7]

Hello @eldxmgw,

Is it possible to share the logs?

Yes, spanning tree something like MSTP.

STP: only supports a single instance whereas MSTP(Multiple) includes the advantages of RSTP, and supports multiple instances.

Is MSTP enabled?

Thanks!

[eldxmgw]

Do you need the buffered log?

I tried to enable differend *STP protocols, but i don't see any change in stable host availability. But i have to make clear i don't really have any clue to configure *STP correctly.

[akg7]

Hello @eldxmgw,

Yes, I need buffer log.

Thanks!

[eldxmgw]

Done via PM.

[akg7]

Hello @eldxmgw 

Thanks received, I will check and let you know.

Thanks!

[akg7]

Hello @eldxmgw,

One more question, what is the current software version in the switch?

Thanks!

[eldxmgw]

I just updated last weekend IMHO to the latest version from 02.16 to 02.18 over this link: https://h10145.www1.hpe.com/downloads/SoftwareReleases.aspx?ProductNumber=JL380A&lang=&cc=&prodSeriesId=&SoftwareReleaseUId=22203&SerialNumber=&PurchaseDate=

[akg7]

Hello @eldxmgw,

Apologies for delayed response.

I have reviewed the log buffer and there are not much information into it which give the reason code.

Also this switch is running on latest version.

The issue needs remote/LAB intervention so I will advise you to log a case with support/

And are you using Aruba central?

Thanks!

---

---

[eldxmgw]

Hello, sorry i don't use any support or similar board concerning HPE stuff than this.

I don't even know where, what and how to do it.

I just register on this community board hoping to finally get this issue fixed

[akg7]

Hello @eldxmgw,,

Let me give one more try, is it possible for you to share network/topology diagram?

Thanks!

[eldxmgw]

Ahm, something like that doesn't exists. I don't even know how to do this right now, sorry.

[eldxmgw]

Any further advice, or is this issue falling asleep now?

[akg7]

Hello @eldxmgw,

Is this switch part of Aruba central?

As mentioned erlier, the issue needs remote/LAB intervention so I will advise you to log a case with support by using below link:

https://support.hpe.com/hpesc/public/home/

Thanks!