unable to access E2810 from the core switch after firmware upgraded

[airwolf09]

I have E5406 uplink port A24 to E2810 port 24 and everything were working properly and the E2810 swith is running version 11.25. I then decide to update the E2810 to the latest firmware 11.48.

run "show spanning-tree" on both switches show the uplink port as follow:

E5406 port A24 shows blocking

E2810 port 24 shows forwarding.

Once the firmware upgraded I reboot the switch but unable to ping E2810 from E5406 or vice versa. if I rolled back to the 11.25 firmware then both swith communicate fined.

has anyone run into this problem with the 11.48 firmware for E2810 switch?

is there a way to solve the issue?

[AAPP Toledo]

Hi airwolf,

Do check if you have one or more config files? Check it with the command "show config files".

This is because when you reboot (boot system flash xxxx) every image (Pri/Sec) is asociated to and config file.

For example (Check below how I have two config files and the first one is associated to Primary and second one to Secondary):

Procurve(config)# sh config files

Configuration files:

id | act pri sec | name
---+-------------+------------------------------------------------
1 | * | config1
2 | * * | config2
3 | |

I don't know if this is the REAL problem/solution... only a possibility ;)

Greetings...

[airwolf09]

i'm running the same config file on both firmware.  as I have the config file to be a default boot.

[AAPP Toledo]

Hi Again,

I think you could try posting here the E2810 configuration (Sh Runn Conf) before and after the Version's upgrade. As there is a saying here .... "Two heads are better than one" :)

Greetings....

[airwolf09]

somehow I cannot attached the file so I pasted here. Below is  E2810 config.

port A24 on E5406 is uplink to Port 24 on E2810.

the same config are are run before and after 2810 upgrade. it ran fined on version 11.25 but not 11.48

---

; J9021A Configuration Editor; Created on release #N.11.25

hostname "Satelite-2810-24G"
max-vlans 256
time timezone -300
time daylight-time-rule Continental-US-and-Canada
console inactivity-timer 60
interface 1
   name "HVAC"
exit
interface 24
   name "5406zl Uplink"
exit
ip default-gateway 10.0.1.1
sntp server 192.5.41.209
timesync sntp
sntp unicast
sntp 30
snmp-server community "public" Unrestricted
vlan 1
   name "NotUsed"
   no ip address
   no untagged 1-24
   exit
vlan 11
   name "ManageIP"
   untagged 1
   ip address 10.0.1.27 255.255.255.0
   tagged 24
   exit
vlan 12
   name "Production2"
   untagged 2-23
   tagged 24
   exit
vlan 13
   name "Development"
   tagged 24
   exit
vlan 14
   name "Internal WAP"
   tagged 24
   exit
vlan 15
   name "GMP"
   tagged 24
   exit
vlan 17
   name "Public WAP"
   tagged 24
   exit
vlan 18
   name "DMZ"
   tagged 24
   exit
vlan 19
   name "MitelVOIP"
   voice
   exit
vlan 20
   name "Backup/VM"
   exit
vlan 21
   name "iSCSI"
   jumbo
   exit
vlan 10
   name "Production"
   exit
primary-vlan 12
management-vlan 11
spanning-tree

[airwolf09]

here is my 5406 config

; J8697A Configuration Editor; Created on release #K.15.05.0002
; Ver #01:01:00

hostname "Metabolon-Satelite-Switch-5406zl"
time timezone -300
time daylight-time-rule Continental-US-and-Canada
module 1 type J8702A
module 2 type J8702A
module 3 type J9309A
module 4 type J9309A
interface A1
   name "IPCAM"
exit
interface A2
   name "IPCAM"
exit
interface A3
   name "IPCAM"
exit
interface A4
   name "WAP"
exit
interface A5
   name "WAP"
exit
interface A24
   name "2810-24G Uplink"
exit
interface C1
   name "Fiber Channel 1"
exit
interface C3
   name "Fiber Channel 2"
exit
trunk C2,C4 Trk1 LACP
trunk C1,C3 Trk2 Trunk
ip default-gateway 10.0.1.1
vlan 1
   name "NotUsed"
   untagged D1-D4,Trk1-Trk2
   no untagged A1-A24,B1-B24
   no ip address
   exit
vlan 11
   name "ManageIP"
   untagged A1-A3,B3
   ip helper-address 10.0.0.9
   ip helper-address 10.0.0.19
   ip helper-address 10.0.0.21
   ip helper-address 10.0.0.23
   ip address 10.0.1.3 255.255.255.0
   tagged A24,Trk1-Trk2
   exit
vlan 10
   name "Production"
   ip helper-address 10.0.0.9
   ip helper-address 10.0.0.19
   ip helper-address 10.0.0.21
   ip helper-address 10.0.0.23
   tagged A24,Trk1-Trk2
   no ip address
   exit
vlan 12
   name "Production2"
   untagged A6-A23,B1-B2,B4-B24
   ip helper-address 10.0.0.9
   ip helper-address 10.0.0.19
   ip helper-address 10.0.0.21
   ip helper-address 10.0.0.23
   tagged A24,Trk1-Trk2
   no ip address
   exit
vlan 13
   name "Development"
   ip helper-address 10.0.0.9
   ip helper-address 10.0.0.19
   ip helper-address 10.0.0.21
   ip helper-address 10.0.0.23
   tagged A24,Trk1-Trk2
   no ip address
   exit
vlan 14
   name "InternalWireless"
   untagged A4-A5
   ip helper-address 10.0.0.9
   ip helper-address 10.0.0.19
   ip helper-address 10.0.0.21
   ip helper-address 10.0.0.23
   tagged A24,Trk1-Trk2
   no ip address
   exit
vlan 15
   name "GMP"
   ip helper-address 10.0.0.9
   ip helper-address 10.0.0.19
   ip helper-address 10.0.0.21
   ip helper-address 10.0.0.23
   tagged A24,Trk1-Trk2
   no ip address
   exit
vlan 17
   name "PublicWireless"
   tagged A4-A5,A24,Trk1-Trk2
   no ip address
   exit
vlan 18
   name "DMZ"
   tagged Trk1-Trk2
   no ip address
   exit
vlan 19
   name "MitelVoip"
   tagged A6-A18,B1-B6,B8,B17-B24,Trk1-Trk2
   voice
   no ip address
   exit
vlan 20
   name "Backup/vMotion"
   tagged A24,Trk1-Trk2
   no ip address
   exit
vlan 21
   name "iSCSI"
   tagged A24,Trk1-Trk2
   no ip address
   jumbo
   exit
console inactivity-timer 60
timesync sntp
sntp unicast
sntp 30
sntp server priority 1 192.5.41.209
ip route 0.0.0.0 0.0.0.0 10.0.0.254
snmp-server community "public" unrestricted
spanning-tree
spanning-tree Trk1 priority 4
spanning-tree Trk2 priority 4
spanning-tree priority 1
primary-vlan 12
management-vlan 11

[AAPP Toledo]

Hi,

I can see that in both switches the ports (A24 and 24) are no untagged on VLAN 1 and tagged on the rest of VLANs but don't UNTAGGED on no-one VLAN.... Why?

Why do you try to configure a TRUNK in both switches? For example....

In E2810

trunk 24 Trk3 LACP

In 5406zl

trunk A24 Trk3 LACP

And obviously change in all VLANS the tagged 24 to tagged Trk3 and tagged A24 to tagged Trk3. Anyway but I think that this change is automatic.

I have 4 5406zl and several 2610/2510 in different floors connected to a 5406zl core, and this is the way (LACP's Trunk) to connect it.

I hope to help you in this odyssey...

Greetings...

[airwolf09]

I don't need trunk uplink on 2810 as i only have 1 uplink cat6 cable. so it doesn't make any sense for trunking 5406 to 2810.

I don't use vlan 1. vlan 12 is untagged if you look closely.

don't think trunking solve the problem of why 11.48 is not working and 11.25 is working.

[AAPP Toledo]

Hi again,

You say that vlan 12 is untagged if you look closely but I can see...

hostname "Satelite-2810-24G"

vlan 12
   name "Production2"
   untagged 2-23
   tagged 24
   exit

hostname "Metabolon-Satelite-Switch-5406zl"

vlan 12
   name "Production2"
   untagged A6-A23,B1-B2,B4-B24
   ip helper-address 10.0.0.9
   ip helper-address 10.0.0.19
   ip helper-address 10.0.0.21
   ip helper-address 10.0.0.23
   tagged A24,Trk1-Trk2
   no ip address
   exit

I don't see that the ports (24 and A24) are untagged.... Why do you test to do a real untagged in both ports?

Anyway... I think if you open a case in HP Support they can say something about it.

Greetings...

[airwolf09]

You can not untagged vlan between switches on the uplink port. Vlav 12 is primary vlan for that swith and it it untagged for port that are not use for uplink between switch. It must be tagged on uplink port.
I already had a case open with HP.

[AAPP Toledo]

Hi Airwolf,

I have several switchs with untagged on uplink ports and running fine. An example...

************************************************************************************************************

SWITCH (1)
SW_Arriba_CPD012# show cdp neighbors detail
  Port : 23 
  Device ID : 00 18 71 90 9c c0                                            
  Address Type : IP         
  Address      : 10.128.24.244                                              
  Platform     : ProCurve J4903A Switch 2824, revision I.10.77, ROM I.08....
  Capability   : Switch                                                    
  Device Port  : 24                                                        
  Version      : ProCurve J4903A Switch 2824, revision I.10.77, ROM I.08....

vlan 1
   name "DEFAULT_VLAN"
   untagged 9
   no ip address
   no untagged 1-8,10-24
   exit
vlan 2
   name "DATOSARCHIVO"
   untagged 1-8,13,18,23-24
   ip address 10.128.24.174 255.255.255.128
   exit

vlan 3
   name "VOZ ARCHIVO"
   untagged 10
   no ip address
   tagged 11-20,23
   voice
   exit

SWITCH (2)

SW_Archivo_Anexo# sh cdp neighbors detail

  Port : 24 
  Device ID : 00 17 08 ad 15 40                                            
  Address Type : IP         
  Address      : 10.128.24.174                                              
  Platform     : ProCurve J4903A Switch 2824, revision I.10.77, ROM I.08....
  Capability   : Switch                                                    
  Device Port  : 23                                                        
  Version      : ProCurve J4903A Switch 2824, revision I.10.77, ROM I.08....

vlan 1
   name "DEFAULT_VLAN"
   no ip address
   no untagged 1-24
   exit
vlan 2
   name "DATOSARCHIVO"
   untagged 1-24
   ip address 10.128.24.244 255.255.255.128
   exit

vlan 3
   name "VOZ ARCHIVO"
   tagged 23-24
   exit

************************************************************************************************************

In both switches the uplink PORT is untagged. On Switch(1) port 24 untagged on VLAN2 and on Switch(2) port 23 untagged on the same VLAN2.... What's the problem?

SW_Arriba_CPD012# sh uptime
0279:22:34:28.67
SW_Archivo_Anexo# sh uptime
0279:22:39:23.88

Greetings....

[Arimo]

Hi

Since STP is blocking the port I'd be surprised if it allowed traffic. Looks like there's been some change in STP calculation between the versions. This might be a result of fixes 57003 or 61650.

If you're running MSTP but haven't defined instances, I'd suggest configuring just RSTP. Then adjust the STP link and root priorities so that you know exactly which links in the network are blocked.

[MikeMay]

did you figure out a fix, this just happened to me and am looking for a reason why

[airwolf09]

no, not yet. i have been working with HP tech try to find out why but they haven't able to figure out.  I have try a bunch of their suggestion and none were working at this time.

they have a bunch of revision between 11.25 to 11.48 but none of them were released excpet 11.25 and 11.48.

[MikeMay]

wow, how long has HP been working on this issue?, this post is from 2009/2010 what is strange 11.15 and 11.25 work without issue, add 11.48 and you have a dead switch

[airwolf09]

have you open a case with them?  if not, try to open a case and so that they get more info to solve the issue.

[david-p]

Hi any news with this, or does the request to HP support need further info? I also upgraded a 2810 to 11.48 and the MST instances no longer see the root switches. I have double checked the configuration, all looks ok. Going back to 11.25 with the same config makes it work again as normal. MST on the 2810 running 11.48 shows every port as edge, which is not correct.

[airwolf09]

nope, they have not come up with a solution yet.