Ping randomly timed out

[securitas057]

Hi

Lately we upgraded our network with 8 new cisco 1300 switches, but not our HP core switch. We have 7 vlans and routing is done on core switch.

127.0.0.0/8 reject static 0 0
127.0.0.1/32 lo0 connected 1 0
172.14.0.0/24 IOT 40 connected 1 0
172.16.1.0/24 DEFAULT_VLAN 1 connected 1 0
192.168.1.0/24 Access 10 connected 1 0
192.168.2.0/24 P2P Cameras 2 connected 1 0
192.168.11.0/24 Lenel Access 11 connected 1 0
192.168.15.0/24 Management 15 connected 1 0
192.168.20.0/24 CCTV 20 connected 1 0
192.168.30.0/24 Cameras 30 connected 1 0

Additionaly all cisco switches have

S 0.0.0.0/0 [1/4] via 192.168.15.254, 22:50:15, vlan 15
C 192.168.15.0/24 is directly connected, vlan 15
C 192.168.255.10/32 is directly connected, loopback 1

All end devices are connected but while pinging them I am randomly getting timed out and losing pockets - average 5% on each vlan.

[pieterh]

what first comes in mind is a spanning-tree configuration mismatch between Cisco and HP switch

[securitas057]

How to make it work correctly? which spanning-tree to choose?

[Jens Albrecht]

For mixed vendor environments MSTP is always the recommened option.

The reason is very simple.
MSTP is the only version that Cisco implemented according to RFC standards and therefore, it is the only option known to work for sure with other vendor switches.

[Joseph W. Doherty]

Since the 1300 switches are L3, and you show a static default route and a loopback (management) in a different network, what's the L2/L3 topology?

Both @pieterh and @Jens Albrecht are correct to consider spanning tree issues, as Cisco, by default, uses PVST, while other vendors, by default, use the standard version.

I recall, but at the moment cannot find, a Cisco document that explains how to interoperate between standard STP (ideally rapid) and Cisco's equivalent PVSTs.  (Not for the faint of heart).  Jens' suggestion to use MST is excellent.

Or, use L3 between the two vendors rather than L2.

[Joseph W. Doherty]

Oh, I also meant to ask, is the only issue appear to be ping?

Reason I ask, the Cisco 1300s might be rate limiting ping.

[securitas057]

All devices are connected and I did not find that devices are getting disconnected.

beside routing table on core switch I have added

S 0.0.0.0/0 [1/4] via 192.168.15.254, 22:50:15, vlan 15

to all cisco switches because I could not connect to them from my workstation which is on vlan 20.

Should I keep it or there is no need for it and I need to find another solution?

[Joseph W. Doherty]

What did you have before the 1300s?

L2 switches, as with other network hosts, to communicate with other hosts, not on the same network, need either a local proxy gateway or a gateway IP.  For a L3 switch or router, they need a route to other networks.  This may explain why you had to add a static route to all the new switches, but it probably doesn't explain the ping issue.

So, you still have the ping issue?  No other apparent operational issues?

[securitas057]

there were Netgears simple dumb switches

[MHM Cisco World]

Each cisco device have control plane protection or CoPP' so ping from device and get 5% drop can from this protections.

To check if your device work good or not try ping through device not using device and source or destination of ping.

MHM

[Joseph W. Doherty]

Each cisco device have control plane protection or CoPP' so ping from device and get 5% drop can from this protections.

A more descriptive and specific variant of my earlier "Reason I ask, the Cisco 1300s might be rate limiting ping."

To check if your device work good or not try ping through device not using device and source or destination of ping.

An excellent suggestion!  For example pinging between edge hosts.

Also it's why I asked about any other operational issues.  I.e. trying to ascertain whether it's a CoPP issue.  Again, though, excellent suggestion.

[securitas057]

Unfortunately

Can't find any link to newest firmware

Your switch is running:

• Software Version: K.16.01.0007 (from 2016)

• Boot ROM: K.15.30

---

What this means for MSTP:

• The HP ProCurve 5412zl with this firmware version typically does NOT support MSTP as a global spanning-tree mode.

• It mainly supports STP and RPVST (Rapid PVST), but not MSTP.

[Jens Albrecht]

Well, the last firmware for these switches was K.16.02.0035 released on Jun 15, 2022 as far as I know.

And according to the official HPE ArubaOS-Switch Advanced Traffic Management Guide for release K.16.02 these switches fully support MSTP. Even much older versions do support it as I configured MSTP on HP ProCurve switches running K.15.x software many years ago.

Of course, you need a HPE account to get access to the above mentioned firmware and documentation.

HTH!

[pieterh]

Cisco switches have a command to enable to write logging to flash besides console/tftp/etc ("logging file ....")
HP procurve may have this option too, so you can retreive the log after a power-cycle.

(syslog does not help because syslog server is not available when network is down)