Weirdly dying HP 2610 switches

[CraigM]

Hi everyone,

 

I've got a stack of old HP 2610 switches in my office. I had a donation request this week, so I pulled 2 of the 24 port models out to clean and wipe the config. Both of the units refuse to boot, they just make a high-pitched squealing noise (very similar to coil whine) and every light on the switch stays lit up. There's no console port access and the switch never goes further.

 

This isn't the first time this has happened to me. Last year I had the same issue with 2 other 2610's I was donating as well. Logged with HP and they replaced the switches as per the warranty but I never found out what the root problem was.

 

The funny thing is that these switches were absolutely bulletproof during their service lives, never had any issues. But now after sitting on a shelf for a year or more after being retired, they just randomly die? I've taken the tops off and I don't see any leaking, bulging or exploded capacitors, but I lack the skill to troubleshoot further.

 

Was just wondering if something like this has happened to anyone else with this particular model? I know electronics age and nothing is completely bulletproof, but this just seems weird to me.

 

Thanks!

[psydii]

Had one die a couple of weeks ago after it got powered down for the first time in a couple of years.

[CraigM]

Had one die a couple of weeks ago after it got powered down for the first time in a couple of years.

 

So weird that being powered down for an extended period causes it to die.

 

I've now logged both switches with HP, will probably get 2 refurb 2610's for replacement if last year was anything to go by.

[dgsmith]

will probably get 2 refurb 2610's for replacement if last year was anything to go by.

 

Depends on stock levels at the time the claim is made. One time I got lucky and they swapped a 2510 for a 2530 as the wait time on a refurb 2510 would have been outside the SLA, though usually as you say, it's like-for-like, not brand new. I have perhaps 2 or 3 26x0 switches in storage for about 2 years, not booted in that time but wouldn't have expected they may not, so I might check these just in case, as they are the emergency swap-out stock!

[MrEprise]

I know you said that you couldn't find any leaking/bulging/exploded capacitors but sometimes they can fail without showing any physical signs of doing so. I've had monitors before where the backlights have dimmed considerably and after replacing what seemed like perfectly healthy caps, this has often fixed the backlight issues.

[CraigM]

Depends on stock levels at the time the claim is made. One time I got lucky and they swapped a 2510 for a 2530 as the wait time on a refurb 2510 would have been outside the SLA, though usually as you say, it's like-for-like, not brand new. I have perhaps 2 or 3 26x0 switches in storage for about 2 years, not booted in that time but wouldn't have expected they may not, so I might check these just in case, as they are the emergency swap-out stock!

 

Good idea to check them out. One never knows. Luckily these 2 units aren't ever going back into my network and I have more units, but this will now make 4 units that have done this to me. Once long ago, we got a 2620 in place of a 2610, but otherwise it's been like for like.

 

I know you said that you couldn't find any leaking/bulging/exploded capacitors but sometimes they can fail without showing any physical signs of doing so. I've had monitors before where the backlights have dimmed considerably and after replacing what seemed like perfectly healthy caps, this has often fixed the backlight issues.

 

Thanks yeah, I do know that they can do that. Unfortunately I'm not set up to test caps (not that I have the soldering skills to apply/remove them in the 1st place ) I've been tempted to maybe test with another 2610 power supply to narrow down whether it's the power supply or the board itself, but now that I've logged the switches I don't want to fiddle further

[PotNoodleTech]

It's a funnny one. Equipment that has been powered 24/7 for over a decade can sometimes not like being unplugged.

 

Its likely they're toast - they must be what 15 years old now?

[Mako]

Can confirm that range of HP switches can just die after a power cycle. Lost about 6 in the last 3 years after various power outages, and one of those was just a case of moving it from one cab to another and it just never came on again.

 

Fortunately the lifetime warranties are a saviour...

[psydii]

Given they are clearly signaling that we should be prepared to manage through 4hr power cuts this winter, I wonder how many other switches are going to fail. I don't think our core switch has powered down in 8 years!

[CraigM]

It's a funnny one. Equipment that has been powered 24/7 for over a decade can sometimes not like being unplugged.

 

Its likely they're toast - they must be what 15 years old now?

 

These 2610's are 11 years old now. We got them about 2 years after I started at this school, when my 1st partner colleague insisted we move to managed switches throughout and get rid of the old unmanaged D-Link switches the school had previously. In operation those 2610's were nigh bulletproof.

 

Can confirm that range of HP switches can just die after a power cycle. Lost about 6 in the last 3 years after various power outages, and one of those was just a case of moving it from one cab to another and it just never came on again.

 

Fortunately the lifetime warranties are a saviour...

 

Thankfully we never had that issue despite having increasingly regular power outages over the last few years here in South Africa - one small mercy. However I cannot agree enough about those lifetime warranties! That alone has inspired such confidence in me to always go with HPE/Aruba.

 

Given they are clearly signaling that we should be prepared to manage through 4hr power cuts this winter, I wonder how many other switches are going to fail. I don't think our core switch has powered down in 8 years!

 

Here in South Africa, regular cuts called load shedding have become the norm over the last few years as the only power company in the country convulses in its death throes. Our equipment regularly gets powered off and thankfully we've had no issues to date. Long term I don't know if the constant cuts are subtly and slowly damaging the power supplies in the current switches, so only time will tell.

[dmj]

If it's been powered off for a long time there will be no electrons flowing to prevent internal corrosion.

Corrosion will build up resistance in components significantly, that resistance will cause failure across a component.

That's why electronics are packaged damp proofed with silica gel packets.

[psydii]

Load Shedding is a term I have heard here. Also Rota disconnection (as in, there is a rota for who loses power when) seems to be the term in the official documentation:

 

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/995049/esec-guidance.pdf

 

A lot of schools here will have 2006-2011 era switches. While post-capacitor plague, electrical kit from this era was more prone to solder fatigue failure. Fatigue being caused by thermal changes (as might be seen when powering down for a few minutes), and corrosion (as might be seen if a device is on the shelf for a while). Solder quality improved as time went on so newer kit is less prone to these failures.