Stacking two HP ProCurve Switch 2824 (J4903A) consideration and discussion ?

[Server-Support]

Hi All,

I'm currently planning to stack two  HP ProCurve Switch 2824 (J4903A) switch in my server room so that I can manage it from single management IP address console (through web based interface).

just like in the page 46 of https://www.manualslib.com/manual/275006/Hp-Procurve-2824.html?page=46

So can anyone here please clarify the consequence of stacking it:

1. If one switch eg. the Commander is down due to power outage or firmware update / maintenance, does the other switch can still be working for anything that connect to it ?

2. Assuming the management IP address of the stacked switches is just one, can I still access the switch individually by the IP address through web browser ? To backup the switch config before the firmware maintenance.

Sorry to ask such silly questions since I'm new to HP networking.

Thanks in advance.

[parnassus]

Never tried "stack for management" but this sentence makes me worrying a little bit in case of Commander goes down for a long time (let's say it faults):

"The management includes Telnet access and web browser interface access to the Commander and to each Member switch through the Commander"

since you have just *one* IP address (the Commander's IP address) to connect to.

Consider those General Stacking Operation notes (not tied to 2800 specifically):

"After you configure one switch to operate as the Commander of a stack, additional switches can join the stack by either automatic or manual methods. After a switch becomes a Member, you can work through the Commander switch to further configure the Member switch as necessary for all of the additional software features available in the switch. The Commander switch serves as the in-band entry point for access to the Member switches. For example, the Commander’s IP address becomes the path to all stack Members and the Commander’s Manager password controls access to all stack Members."

So if the Commander is down (no matter why and for how long) you will be unable to access the Member switches of the same management stack during the whole duration of its unavailability...AFAIK there is no (automatic/manual) mechanism to elect another Commander switch (among the Member of the same stack) if the original Commander is faulted/down...hope not to be wrong here (specific documentation about 2800 could be more detailed about pro/cons of deploying Stack for Management).

[Arimo]

Short and sweet - yes and yes.

As described already, the commander is just a central access location for the stack members. Commander going down means you can't use the commander's IP to access the members, nothing else. If each switch has its own individual IP address, you will be able to access them individually using that.

What comes to web interface; I'd say don't. ProCurves have multiple command interfaces, but the only one which allows you to access all the configuration options is the CLI. All other interfaces have limitations which can sometimes be surprising. Both GUI and Menu interface have their uses (GUI for a status check, Menu for quick'n'dirty configurations) but to properly manage the boxes you should get comfy with the CLI.