Web and Unmanaged
1752286 Members
4765 Online
108786 Solutions
New Discussion юеВ

Procurve 1920-24G Poe - Crashes , Disconnections Weird Issues

 
Kleanthis
Occasional Visitor

Procurve 1920-24G Poe - Crashes , Disconnections Weird Issues

Ok so recently we have installed two new Procurve 1920-24G poe switches to our companies 2nd floor. The first switch was connected directly to an empty slot on a Checkpoint Firewall and and the 2nd was connected to the 1st switch . On the first day everything was working fine , just plug and play no problem . 2nd day we come in the office there is no network at all. The switch was not regignising the connection from the checkpoint firewall and therefore there was no network. When we connected a pc to the cable that was coming from Checkpoint it connected fine but the switch wouldnt . We tested several cables , ports on the switch , we tested going via another managed switch to the procurve switch but nothing. The only thing that wokred was placing a dummy switch in betweek checkpoint and 1920-24G . This is the only was the switch can connect to our lan . Maybe it is something that has to do with the speed ? Downing the speed from Gigabit fot 100mbit fixes the issue? But why is that i dont get it . Additionally after approximately 3 weeks of working with the dummy switch as intermediate the switch crashes on a non working day (ver low network usage ) .and has to be restarted twice for it to work . could this just be a faulty switch issue or there something else here ?

5 REPLIES 5
parnassus
Honored Contributor

Re: Procurve 1920-24G Poe - Crashes , Disconnections Weird Issues

What's about going a little bit more deep and check (both) 1920's logs? I presume you setup them both with IP Management address (that's on Default VLAN, generally).

What's about providing informations about (a) 1920 Switches running software versions and actual running configurations (to diagnose what's happening on the CheckPoint Firewall <--> 1920 link, first of all), (b) CheckPoint connected port settings and (c) how did you setup the link between both 1920s (I imagine you configured both involved ports as Port Type "Trunk" since both ports are intended to interconnect Switches and not edge devices like happen on Port Type "Access" set ports).

Generally HPE OfficeConnect 1920 ports default setting is set to "Auto" (for Port Speed recognition), that means the Switch tries auto-negotiation with remote peer (host, in your case the Firewall) on each port.

What's about also checking Duplex Mode (Full-/Half-Duplex) on all involved ends?

The HPE OfficeConnect 1920 Switch series is pretty stable if correctly configured...

Edit: this post should be definitley moved into "Comware based" Switches category or "Web and Unmanaged" (since it belongs to Smart Managed switches familiy) because HPE OfficeConnect 1920 Switch series is not based on HP ProVision OS (used on old HP ProCurve and into new Aruba, rebranded into ArubaOS-Switch).


I'm not an HPE Employee
Kudos and Accepted Solution banner
Kleanthis
Occasional Visitor

Re: Procurve 1920-24G Poe - Crashes , Disconnections Weird Issues

So here is my 1st Switch Info that is connected to Checkpoint via a dummy switch .

Device Name
HPE 1920-24G-PoE+ (180W) Switch JG925A

 

Product Information
HPE 1920-24G-PoE+ (180W) Switch Software Version Release 1115

 

Device Location
2ND FLOOR

 

Contact Information
Hewlett Packard Enterprise Company 3000 Hanover St Palo Alto, CA 94304

 

SerialNum
CN6BGP524H

 

Software Version
5.20.99 Release 1115

 

Hardware Version
REV.A

 

Bootrom Version

117

 

Running Time:
0 days 3 hours 43 minutes 9 seconds

 

I didnt touch any of the settings i left them all to defaults as i have no vlans that will pass via this switch just the main one .

I have no logs of the switch as i have not set any syslog server yet . 

 

The ports are all setup as access but i dont see why would this matter since its working if i put a dummy switch in between also , on the first day of the setup it worked just fine . Speed is set to AUTO (100mbit) since the dummy switch is 100mbit max. Duplex mode is set to AUTO on all swicthes ports. 

parnassus
Honored Contributor

Re: Procurve 1920-24G Poe - Crashes , Disconnections Weird Issues


Kleanthis wrote: I have no logs of the switch as i have not set any syslog server yet .

That doesn't mean nothing: you don't need an external Syslog Server to see the Switch logs. On Web Interface you can see (some) logs already (as example to see if there is a port which is flapping)...then you can check Port Status and Port Statistic of port facing your CheckPoint Firewall to learn if there is something abnormal on traffic counters (look for Tx/Rx Errors)...or you can try diagnostic tools (Ping/Traceroute) too.

Since PoE is enabled by default on all ports, have you just tried to set (PoE->Port Setup) the "Power State" to "Disable" on the Port used to link to CheckPoint Firewall?

Is the cabling (UTP/STP patch cords) you use Cat.6 (or, at worst, Cat.5E) certified?

The software version appears updated: R1115 is the latest one available.


I'm not an HPE Employee
Kudos and Accepted Solution banner
Kleanthis
Occasional Visitor

Re: Procurve 1920-24G Poe - Crashes , Disconnections Weird Issues

I have used another cable connected from another switch closer than before . Factory made cable and it works. Here is the paradox:

 

1. Before me there was another company which was using the same cabling that i have tried using without success and they had no issues. 

2. I have tried with another cable the factory made one connected to a switch closer to my switch instead of the checkpoint and it works fine Gigabit and all.

3. When i connect a pc to the old cabling it works fine on some pcs on some not but never on the switch .

 

What is happening , i am thinking its probably the cable but why does it work when connected directly to a pc then ?

parnassus
Honored Contributor

Re: Procurve 1920-24G Poe - Crashes , Disconnections Weird Issues

Passive cabling check is a must: I mean, you need to be 100% sure your cabling end-to-end (patch cables included, on both ends) are certified for working under Cat.5E or Cat.6 standards (prefer Cat.6 if you plan to work on a fully Gigabit Ethernet enabled network). Sometime if you don't touch a patch cable or a RJ45 wall outlet for years it just works then, suddenly after you touch it, it could start to behave weirdly (copper contacts? bad punched wires? who knows?).

If the network is (and so was) certified for Cat.5E or Cat.6 you have less troubles because the certification involved a stric testing on cables (under 90 meters), wall outlets, patch cables, RJ45 rack panels and so on.

To troubleshoot (AKA eliminate) a factor from the equation...just connect directly (as you did, right?) your apparatus (if you can)...and then test...that way your only passive cabling factor is the patch cable between your apparatus and you can easily and quickly recognize if something is wrong or not with it (or with Switches' ports connected thorough it).


I'm not an HPE Employee
Kudos and Accepted Solution banner