2530-48G vlan/link aggregation config troubles

[Fred_PyradiaInc]

Hello. Here's my issue.

I have three 48-port switches that my company plan to use as replacements for our present switch setup. I'm trying to set them up to achieve two things:

• Have separate VLANs for each department (admins, sales, purchases, etc...) whom will be able to share the same resources (i.e.: a printer VLAN).
• In order to increase bandwidth, I want to use link aggregation to connect Switches 1&2 with 2 1gb ethernet cables, same between Switches 2&3, and Switches 1&3. I figured I'd also connect each switch to the router too, so I'm going to be using 7 ports total on each switch to achieve all that redundant inter-connectivity.

Now, I have some basic familiarity with VLAN/trunking, but it comes from following Cisco-given classes; HP switches do things differently, and despite browsing forums to glean some info, I end up very confused (some terms mean the reverse of what I've been taught they mean), and not at all confident of what I'm doing.

I could use certainly use some handholding here.

Right now, I have two of those new switches out of the box and plugged so I can run simulations on them and figure them out before I go configure them to meet my company's needs. Here's how what their "running config" looks like right now:

hostname "Switch1"
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   no untagged 13-24
   untagged 1-12,25-52
   ip address dhcp-bootp
   exit
vlan 10
   name "admin"
   untagged 13-24
   tagged 1-12
   no ip address
   exit
vlan 99
   name "management"
   tagged 1-12
   ip address 192.168.0.1 255.255.255.0
   exit

hostname "Switch2"
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   no untagged 13-24
   untagged 1-12,25-52
   ip address dhcp-bootp
   exit
vlan 10
   name "admin"
   untagged 13-24
   tagged 1-12
   no ip address
   exit
vlan 99
   name "management"
   tagged 1-12
   ip address 192.168.0.2 255.255.255.0
   exit

So, the objective here was to get ports 1 through 12 to do my communications between switches/router, my 'trunking' ports. That's apparently achieved by tagging them in each VLAN, meaning thatmy VLAN agrees to share it with other VLANs.

I've a "Management" VLAN, which was made so that I could put an address there that I'd use for each Switch to ping each other.

Ports 13 through 24 were to be used only by my "Admin" VLAN. Untagged apparently means that I assign them to that VLAN, but that those won't be shared under any circumstances with other VLANs.

I'm mostly ignoring DEFAULT_VLAN for now. Honestly, I don't know what to do with it anyways and it's apparently not secure to put any configurations there.

...

So, as is, the 'trunking' seems to be working so far. My ping can pass from Switch1 to Switch2 and vice-versa through Switch 1 port 1 to Switch 2 port 1 with the "Management" vlan addresses. However, I've no inkling of if plugging a second ethernet cable (Switch1int2 to Switch2int2) would permit my communications between the two switches a greater throughput.

[Vince_Whirlwind]

As you have figured out,

Cisco "Access" port = Procurve "Untagged"

Cisco "Trunk" port = Procurve "Tagged" multiple VLANs

Also, Cisco "Native" VLAN = "Procurve "Untagged".

Also, when you create a Cisco "Trunk" port, by default all VLANs are allowed. On Procurve, you have to explicitly add VLANs as "Tagged" one by one.

Link aggregation. On the Procurve you configure globally, "trunk 1,2 trk1 trunk". Perform this on both switches, and you have aggregated these two links, it creates a new interface, called "Trk1". You have to configure the Trk1 interface, just like you have to configure the Po1 interface for a Cisco Etherchannel.

[Fred_PyradiaInc]

Thank you for replying. :)

I entered the "trunk 1,2 trk1 trunk" command on both my switches. I did wonder why it wasn't "trunk 1,2 trk1 lacp", as I thought LACP stood for "Link Aggregation Control Protocol".

Following that, I attempted my pings - no success there.

When I looked at my configuration, I realized that the tagged entries for each of my VLANs (10 admin, 99 management) were "3-12" rather than the "1-12" I had originally. Seeing how the default_vlan added trk1 to itself, I figured I needed to add trk1 to the tagged port of VLANs 10 and 99.

I attempted new pings then, but got 'request timed out' on my attempts.

  Switch2(config)# ping 192.168.0.1
  Request timed out.

Here's how it looks on both switches right now:

hostname "Switch1"
trunk 1-2 trk1 trunk
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   no untagged 13-24
   untagged 3-12,25-52,Trk1
   ip address dhcp-bootp
   exit
vlan 10
   name "admin"
   untagged 13-24
   tagged 3-12,Trk1
   no ip address
   exit
vlan 99
   name "management"
   tagged 3-12,Trk1
   ip address 192.168.0.1 255.255.255.0
   exit
spanning-tree Trk1 priority 4

hostname "Switch2"
trunk 1-2 trk1 trunk
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   no untagged 13-24
   untagged 3-12,25-52,Trk1
   ip address dhcp-bootp
   exit
vlan 10
   name "admin"
   untagged 13-24
   tagged 3-12,Trk1
   no ip address
   exit
vlan 99
   name "management"
   tagged 3-12,Trk1
   ip address 192.168.0.2 255.255.255.0
   exit
spanning-tree Trk1 priority 4

Since I have spanning-tree listed, I'm assuming I'm not getting timed-out requests on my pings due to a broadcast storm.

That said, I really don't get why the pings don't go through now. What am I missing?

[Vince_Whirlwind]

That config looks OK, but what are you pinging, source/destination, what are their IP addresses, and what ports are they patched to?

[Fred_PyradiaInc]

Okay, physcially I have two HP 2530-48G switches, one atop the other. Both are powered up, and have loaded the configurations I listed above.

Two 2' ethernet cables run from one switch to the other, one is plugged from Switch1 interface 1 to Switch 2 interface 1, the other goes from Switch 1 interface 2 to Switch 2 interface 2.

When I ping from Switch1, I try to contact the ip address of the management VLAN of Switch 2 - that's 192.168.0.2 - and from Switch2 tries to reach the ip address of Switch1's management VLAN - 192.168.0.1

If Switch1 pings herself, it works well enough.

 Switch1(config)# ping 192.168.0.1
 192.168.0.1 is alive, time = 1 ms
 Switch1(config)# ping 127.0.0.1
 127.0.0.1 is alive, time = 1 ms

But when Switch1 tries to reach Switch2's management VLAN's ip address...

 Switch1(config)# ping 192.168.0.2
 Request timed out.

The same goes for Switch2:

 Switch2(config)# ping 127.0.0.1
 127.0.0.1 is alive, time = 1 ms
 Switch2(config)# ping 192.168.0.2
 192.168.0.2 is alive, time = 1 ms
 Switch2(config)# ping 192.168.0.1
 Request timed out.

However, if I unplug one of my ethernet cables and have it run from Switch1 interface 3 to Switch2 interface3 instead, and attempt another ping...

 Switch2(config)# ping 192.168.0.1
 192.168.0.1 is alive, time = 13 ms

It works. So, I have no problem with my tagged ports. The issue is solely with the trunked ports.

So, if I review my running config...

hostname "Switch2"
trunk 1-2 trk1 trunk
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   no untagged 13-24
   untagged 3-12,25-52,Trk1
   ip address dhcp-bootp
   exit
vlan 10
   name "admin"
   untagged 13-24
   tagged 3-12,Trk1
   no ip address
   exit
vlan 99
   name "management"
   tagged 3-12,Trk1
   ip address 192.168.0.2 255.255.255.0
   exit
spanning-tree Trk1 priority 4

I'd figure my stumbling point would possibly be where I bolded. But I've no inkling as to how or why, or what I'd be supposed to write to have the desired results.

[Vince_Whirlwind]

Looks fine to me.

Do you need "management vlan 99" configured on each switch?

You should start with "show lldp inf rem" and see if the trunk gives you layer-2 connectivity.

What does show trunk give you?

[Fred_PyradiaInc]

Thanks Vince, I appreciate the help. Hopefully, you'll forgive the growing pains of my fledgling comprehension of the CLI in HP switches.

"show lldp inf rem" only gave me the information for one port, when I'm pretty sure it should've been 2. So, I have layer 2 connectivity, but obviously something was up.

So, I kind of unplugged everything, reset back to factory default, and took a step back to try this again from a fresh start with what I learned thus far:

1. The Management VLAN was something I learned when studying Cisco switches and routers, which made it so that I could give a singular ip address to a switch. In hindsight, that doesn't seem to have worked very well here.

2. Based on our needs, it doesn't feel like we really need multiple VLANs. What I needed was to isolate the heavy internal bandwidth use from mechanical engineering users from those whom made lighter use of the network like the administration.

 Using VLANs would help with broadcasts, but we're a small company so the gains from that are minimal. And it's not going to help with networking resources since the switch is limited to a throughput of 4 gigs.
 The solution seemed to lie in simply putting my different user departments on different switches, rather tham different VLANs.

3. So, I decided to stick with only one vlan, DEFAULT_VLAN.  I also added a third switch. Here's my present setup:

hostname "Switch1"
trunk 1-2 trk1 trunk
trunk 5-6 trk3 trunk
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   untagged 3-4,7-52,Trk1,Trk3
   ip address 192.168.0.1 255.255.255.0
   exit
spanning-tree Trk1 priority 4
spanning-tree Trk3 priority 4

hostname "Switch2"
trunk 1-2 trk1 trunk
trunk 3-4 trk2 trunk
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   untagged 5-52,Trk1-Trk2
   ip address 192.168.0.2 255.255.255.0
   exit
spanning-tree Trk1 priority 4
spanning-tree Trk2 priority 4

hostname "Switch3"
trunk 3-4 trk2 trunk
trunk 5-6 trk3 trunk
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   untagged 7-52,Trk1-Trk3
   ip address 192.168.0.3 255.255.255.0
   exit
spanning-tree Trk1 priority 4
spanning-tree Trk2 priority 4
spanning-tree Trk3 priority 4

Running the "show lldp inf rem" command on Switch1 gives me this...

 LLDP Remote Devices Information

  LocalPort | ChassisId                 PortId PortDescr SysName
  --------- + ------------------------- ------ --------- ----------------------
  1         | 74 46 a0 e2 77 c0         1      1         Switch2
  2         | 74 46 a0 e2 77 c0         2      2         Switch2
  5         | 74 46 a0 e2 59 c0         5      5         Switch3
  6         | 74 46 a0 e2 59 c0         6      6         Switch3

...which is more than I got earlier.

"Show Trunks" gives me this:

 Load Balancing Method: L3-based (Default), L2-based if non-IP traffic

  Port | Name                             Type      | Group Type
  ---- + -------------------------------- --------- + ----- --------
  1    |                                  100/1000T | Trk1  Trunk
  2    |                                  100/1000T | Trk1  Trunk
  5    |                                  100/1000T | Trk3  Trunk
  6    |                                  100/1000T | Trk3  Trunk

 Tested my pings. Got significantly better results, but not without some hiccups:

 Switch1(config)# ping 192.168.0.1
 192.168.0.1 is alive, time = 1 ms
 Switch1(config)# ping 192.168.0.2
 Request timed out.
 Switch1(config)# ping 192.168.0.2
 192.168.0.2 is alive, time = 21 ms
 Switch1(config)# ping 192.168.0.3
 Request timed out.
 Switch1(config)# ping 192.168.0.3
 192.168.0.3 is alive, time = 31 ms

So, switch1 works, sometimes.
 Here's anothe sample from Switch2:

 Switch2(config)# ping 192.168.0.1
 Request timed out.
 Switch2(config)# ping 192.168.0.1
 Request timed out.
 Switch2(config)# ping 192.168.0.1
 Request timed out.
 Switch2(config)# ping 192.168.0.1
 Request timed out.
 Switch2(config)# ping 192.168.0.1
 Request timed out.
 Switch2(config)# ping 192.168.0.1
 Request timed out.
 Switch2(config)# ping 192.168.0.2
 192.168.0.2 is alive, time = 2 ms
 Switch2(config)# ping 192.168.0.3
 Request timed out.
 Switch2(config)# ping 192.168.0.3
 Request timed out.
 Switch2(config)# ping 192.168.0.3
 Request timed out.
 Switch2(config)# ping 192.168.0.3
 Request timed out.

See, before I got to writing this post, I tested pings on all three switches and it was the one that gave me no problems initially. Switch3 could ping the two others fine. It was switch1 that gave me trouble. But just now, though? All request timed out from Switch2.

Ping sample from Switch3...

Switch3(config)# ping 192.168.0.1
Request timed out.
Switch3(config)# ping 192.168.0.2
Request timed out.
Switch3(config)# ping 192.168.0.3
192.168.0.3 is alive, time = 1 ms
Switch3(config)# ping 192.168.0.2
Request timed out.
Switch3(config)# ping 192.168.0.2
Request timed out.
Switch3(config)# ping 192.168.0.1
192.168.0.1 is alive, time = 12 ms

It kind of works, and kind of doesn't.

Returning to ping on Switch1 for kicks, and...

Switch1(config)# ping 192.168.0.1
192.168.0.1 is alive, time = 1 ms
Switch1(config)# ping 192.168.0.2
192.168.0.2 is alive, time = 8 ms
Switch1(config)# ping 192.168.0.3
192.168.0.3 is alive, time = 18 ms
Switch1(config)# ping 192.168.0.2
Request timed out.
Switch1(config)# ping 192.168.0.3
Request timed out.

It can't seem to figure out if it wants to work, or not.

Now, the whole point of using trunks is to benefit from link aggregation and have twice the transfer rate between switches than I'd have usually with a 1gig ethernet cable. But all of these request timeouts kind of defeat the purpose of getting fast connections (and here, I'm just using 3 switches, with nothing else on that network).

So, I figured out using the Management VLAN was my problem before, and I couldn't get any pings going adequately between my management VLAN ip addresses. But now that I'm using the DEFAULT_VLAN, I have something, but it's not reliable.

I'd think Spanning-tree might be the problem... but I need spanning-tree to avoid broadcast storms (besides, once you have trunks, the HP switch doesn't even let you remove spanning-tree anyways). I'm not sure of what else could be done to improve reliability. Any ideas?

[Vince_Whirlwind]

Hmmm...it's like the trunk isn't working. Kind of weird.

You've stablished you have LLDP frames succeeding across the links.

Now check your mac-address tables.

After you ping an address, check the ARP table, see what MAC it's got, then check your mac-address table to see where each switch thinks it is. Look for anything unusual.

Configure each switch with a different spanning-tree priority - make one 4096, another 8192, the last 12288.

show spanning-tree to see which trunk is blocking.

You're definitely going abou this the right way. Build it, break it, learn it.

[Fred_PyradiaInc]

So,I spent the last couple of days experimenting and trying to make heads or tails of my configuration. Going to the point where I eventually just pulled it down to rebuild it. I did not make much headway, though, mostly due to some information just feeling like gibberish to me.

I did confirm that my pings worked perfectly well, until I add the third trunk. Then sometimes they work, sometimes not.

Physically, here are how the switches are set up now.

Switch1 interfaces 1 and 2 connect to Switch2 interfaces 1 and 2. This is Trk1.

Switch1 interfaces 3 and 4 connect to Switch3 interfaces 1 and 2. This is Trk2.

Switch2 interfaces 3 and 4 connect to Switch3 interfaces 3 and 4. This is Trk3.

(I was 3 ports wide on each switch, and now I'm only 2 ports wide)

Here are my running configs:

hostname "Switch1"
trunk 1-2 trk1 trunk
trunk 3-4 trk2 trunk
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   untagged 5-52,Trk1-Trk2
   ip address 192.168.0.1 255.255.255.0
   exit
spanning-tree Trk1 priority 1
spanning-tree Trk2 priority 1

hostname "Switch2"
trunk 1-2 trk1 trunk
trunk 3-4 trk3 trunk
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   untagged 5-52,Trk1,Trk3
   ip address 192.168.0.2 255.255.255.0
   exit
spanning-tree Trk1 priority 2
spanning-tree Trk3 priority 2

hostname "Switch3"
trunk 1-2 trk2 trunk
trunk 3-4 trk3 trunk
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   untagged 5-52,Trk2-Trk3
   ip address 192.168.0.3 255.255.255.0
   exit
spanning-tree Trk2 priority 3
spanning-tree Trk3 priority 3

So, like I said, the pings function as long asmy switches are in a cascading setup. The moment I established Trk3, my pings became unreliable. I also changed my spanning-tree priorities as suggested, but saw no improvement.

Investigating Mac- addresses:

Switch1(config)# show mac-address

  MAC Address   Port   VLAN
  ------------- ------ ----
  7446a0-e259c0 Trk1   1
  7446a0-e259fe Trk2   1
  7446a0-e259ff Trk2   1
  7446a0-e277c0 Trk1   1
  7446a0-e277fe Trk1   1
  7446a0-e277ff Trk1   1

Switch2(config)# show mac-address

  MAC Address   Port   VLAN
  ------------- ------ ----
  7446a0-e259c0 Trk3   1
  7446a0-e259fc Trk3   1
  7446a0-e259fd Trk3   1
  7446a0-e277c0 Trk1   1
  7446a0-e4e0c0 Trk1   1
  7446a0-e4e0fe Trk1   1
  7446a0-e4e0ff Trk1   1

Switch3(config)# show mac-address

  MAC Address   Port   VLAN
  ------------- ------ ----
  7446a0-e259c0 Trk3   1
  7446a0-e277c0 Trk2   1
  7446a0-e277fc Trk3   1
  7446a0-e277fd Trk3   1
  7446a0-e4e0c0 Trk2   1
  7446a0-e4e0fc Trk2   1
  7446a0-e4e0fd Trk2   1

I don't get much of this, except that through my trunks (the appropriate ones, looks like) I get a return from several different MAC address. I thought I'd only get 3 different values total (3 switches), but I'm getting much more.

Now for ARP tables:

Switch1(config)# show arp

  IP Address       MAC Address       Type    Port
  ---------------  ----------------- ------- ----
  192.168.0.2      7446a0-e277c0     dynamic Trk1
  192.168.0.3      7446a0-e259c0     dynamic Trk1

Switch2(config)# show arp

  IP Address       MAC Address       Type    Port
  ---------------  ----------------- ------- ----
  192.168.0.1      7446a0-e4e0c0     dynamic Trk3
  192.168.0.3      7446a0-e259c0     dynamic Trk3

Switch3(config)# show arp

  IP Address       MAC Address       Type    Port
  ---------------  ----------------- ------- ----
  192.168.0.1      7446a0-e4e0c0     dynamic Trk2
  192.168.0.2      7446a0-e277c0     dynamic Trk2

Um... the way I see this...

Switch1 ==> Switch2 ==> Switch3 ==> Switch1

It's a circle loop through all my switches.

As far as I understand, I'm supposed to avoid looping between two switches when trunking and using Spanning-tree, which seems to be the case here. Even though I might have liked it using a more direct path. Maybe I'm missing it, but I'm still not seeing the problem.

On to looking at Spanning-tree:

Switch1(config)# show spanning-tree

 Multiple Spanning Tree (MST) Information

  STP Enabled   : No

...wait, what?
 It's written plainly in my running-config, and it's not actually 'on'?

Switch2(config)# show spanning-tree

 Multiple Spanning Tree (MST) Information

  STP Enabled   : No

Switch3(config)# show spanning-tree

 Multiple Spanning Tree (MST) Information

  STP Enabled   : No

Same on the two others!

...okay, so, how to I enable it? Oh, just "spanning-tree"?

Okay, this is the difference 'enabling' spanning-tree makes in the Running Config:

hostname "Switch1"
trunk 1-2 trk1 trunk
trunk 3-4 trk2 trunk
snmp-server community "public" unrestricted
vlan 1
   name "DEFAULT_VLAN"
   untagged 5-52,Trk1-Trk2
   ip address 192.168.0.1 255.255.255.0
   exit
spanning-tree
spanning-tree Trk1 priority 1
spanning-tree Trk2 priority 1

I did not know that.

Okay, so now, I get something very different from 'show spanning-tree'.

Switch1(config)# show spanning-tree

 Multiple Spanning Tree (MST) Information

  STP Enabled   : Yes
  Force Version : MSTP-operation
  IST Mapped VLANs : 1-4094
  Switch MAC Address : 7446a0-e4e0c0
  Switch Priority    : 32768
  Max Age  : 20
  Max Hops : 20
  Forward Delay : 15

  Topology Change Count  : 3
  Time Since Last Change : 2 mins

  CST Root MAC Address : 7446a0-e259c0
  CST Root Priority    : 32768
  CST Root Path Cost   : 20000
  CST Root Port        : Trk2

  IST Regional Root MAC Address : 7446a0-e4e0c0
  IST Regional Root Priority    : 32768
  IST Regional Root Path Cost   : 0
  IST Remaining Hops            : 20

  Root Guard Ports     :
  Loop Guard Ports     :
  TCN Guard Ports      :
  BPDU Protected Ports :
  BPDU Filtered Ports  :
  PVST Protected Ports :
  PVST Filtered Ports  :

                  |           Prio              | Designated    Hello
  Port  Type      | Cost      rity State        | Bridge        Time PtP Edge
  ----- --------- + --------- ---- ------------ + ------------- ---- --- ----
  5     100/1000T | Auto      128  Disabled     |
  6     100/1000T | Auto      128  Disabled     |
  7     100/1000T | Auto      128  Disabled     |
  8     100/1000T | Auto      128  Disabled     |
  9     100/1000T | Auto      128  Disabled     |
  10    100/1000T | Auto      128  Disabled     |
  11    100/1000T | Auto      128  Disabled     |
  12    100/1000T | Auto      128  Disabled     |
  13    100/1000T | Auto      128  Disabled     |
  14    100/1000T | Auto      128  Disabled     |
  15    100/1000T | Auto      128  Disabled     |
  16    100/1000T | Auto      128  Disabled     |
  17    100/1000T | Auto      128  Disabled     |
  18    100/1000T | Auto      128  Disabled     |
  19    100/1000T | Auto      128  Disabled     |
  20    100/1000T | Auto      128  Disabled     |
  21    100/1000T | Auto      128  Disabled     |
  22    100/1000T | Auto      128  Disabled     |
  23    100/1000T | Auto      128  Disabled     |
  24    100/1000T | Auto      128  Disabled     |
  25    100/1000T | Auto      128  Disabled     |
  26    100/1000T | Auto      128  Disabled     |
  27    100/1000T | Auto      128  Disabled     |
  28    100/1000T | Auto      128  Disabled     |
  29    100/1000T | Auto      128  Disabled     |
  30    100/1000T | Auto      128  Disabled     |
  31    100/1000T | Auto      128  Disabled     |
  32    100/1000T | Auto      128  Disabled     |
  33    100/1000T | Auto      128  Disabled     |
  34    100/1000T | Auto      128  Disabled     |
  35    100/1000T | Auto      128  Disabled     |
  36    100/1000T | Auto      128  Disabled     |
  37    100/1000T | Auto      128  Disabled     |
  38    100/1000T | Auto      128  Disabled     |
  39    100/1000T | Auto      128  Disabled     |
  40    100/1000T | Auto      128  Disabled     |
  41    100/1000T | Auto      128  Disabled     |
  42    100/1000T | Auto      128  Disabled     |
  43    100/1000T | Auto      128  Disabled     |
  44    100/1000T | Auto      128  Disabled     |
  45    100/1000T | Auto      128  Disabled     |
  46    100/1000T | Auto      128  Disabled     |
  47    100/1000T | Auto      128  Disabled     |
  48    100/1000T | Auto      128  Disabled     |
  49              | Auto      128  Disabled     |
  50              | Auto      128  Disabled     |
  51              | Auto      128  Disabled     |
  52              | Auto      128  Disabled     |
  Trk1            | 20000     16   Blocking     | 7446a0-e277c0 2    Yes No
  Trk2            | 20000     16   Forwarding   | 7446a0-e259c0 2    Yes No

Okay, I won't pretend I understand all of this, but it's pretty obvious at the end that Switch1 chose to block Trk1 in favor of Trk2.

I'm not going to bore you with more blocks of text, and will instead just see if my pings work.

Switch1(config)# ping 192.168.0.1
192.168.0.1 is alive, time = 1 ms
Switch1(config)# ping 192.168.0.2
192.168.0.2 is alive, time = 3 ms
Switch1(config)# ping 192.168.0.3
192.168.0.3 is alive, time = 4 ms
Switch1(config)# ping 192.168.0.3
192.168.0.3 is alive, time = 2 ms
Switch1(config)# ping 192.168.0.2
192.168.0.2 is alive, time = 2 ms

Switch2(config)# ping 192.168.0.1
192.168.0.1 is alive, time = 3 ms
Switch2(config)# ping 192.168.0.2
192.168.0.2 is alive, time = 1 ms
Switch2(config)# ping 192.168.0.3
192.168.0.3 is alive, time = 4 ms

Switch3(config)# ping 192.168.0.1
192.168.0.1 is alive, time = 2 ms
Switch3(config)# ping 192.168.0.2
192.168.0.2 is alive, time = 3 ms

No problem here. Pinging works. It was just that missing line about activating spanning-tree that was the problem.

* * *

Whew. Managing to solve that is a huge load off my back.  Thanks for dropping the breadcrumbs leading me to managing a functional trunk system inbetween my two switches.

...I'm not quite done, though. There was something I was trying to achieve - I'd just like to be clear on it.

Back to my objective of using link aggregation to greatly increase my company's intranet speed: I know my 2530 switches have a throughput of 4 gigs. Each ethernet 100/1000T ethernet cable offers 1 gig of bandwidth. If Switch1 ever needed to transmit at a speed of 4 gigs to Switch2, I thought it'd be able to do so using the direct Trk1 for the first 2 gigs, and the indirect Trk3-Trk2 route for the other 2 gigs.

With spanning-tree blocking one trunk out of two on each switch, this seems unfeasible. Is my only way of achieving a 4 gig bandwidth speed only by using a 4-port 100/1000T trunk? It's... physically a lot clunkier than I thought it'd have to be.

[Vince_Whirlwind]

There are two things I would consider if I were in your position:

 - buy switches with 10Gb uplinks

and/or

 - buy switches that "stack" to form a single logical switch

Generally, I create a "core" which has the performance required to carry all the network traffic, then I create an "access" layer of switches that have the performance required to support 24 or 48 hosts, provide PoE, QoS, etc..., and pass all traffic up to the "core".

You appear to be looking at more of a "mesh" situation. You could also buy switches that support a "mesh" - this keeps all links alive instead of using spanning-tree to block them. I would never myself procure switches with the intention of doing this, though.