Finally escaped the 6-month capacity queue

Not naming the facility, sorry. Learned that lesson when I named a provider in 2019 and their "community manager" rang my mobile for three weeks.

I will say it's not Equinix. Their queue was worse. Sixteen weeks quoted for a 3kVA single feed when I checked last autumn.

Own ASN, yes. RIPE allocation, been holding it since 2014. RPKI is done, ROAs published, but you're right about the IX filters -- LINX Manchester took four business days last time I added a new origin. I submitted the LOA yesterday afternoon.

It never does. I have 14 providers in my sheet and only two even publish queue position. The rest use "estimated availability" which is marketing speak for "we'll call you when the sales target is met."

I can share the full methodology if anyone wants. The spreadsheet tracks quoted vs. actual delivery, commit level, and whether the contract includes a "material change" clause that lets them bump you for better terms.

The cash miner phenomenon is in my data. Three facilities, all in North America, showed 40%+ queue compression when crypto prices spiked. Not illegal. Just not FIFO.

Smart. In Vietnam we have same problem. I ask provider about diverse path, they say "yes yes" but same building entry. I bring my own fiber tester now, check myself.

Your LINX Manchester, they use switch or router at port? Some IX still use switch only, no BGP security feature. I hope you check this.

Please do. We're looking at a second site in Toronto and the queue games are worse than Vancouver was three years ago. One provider quoted us February, then June, then "Q3 best effort." We walked.

Also re: crypto -- the Ontario government is talking about a similar 5kW registration rule like Vietnam. Not law yet, but the datacenters are already asking more questions.

This is good, very good. The young engineers, they trust too much the paper. I see "diverse path" that goes same conduit 200 meter, then separate. Is not diverse, no?

Forty years I check with the eyes and the hands. The terminal, it lies nice. The cable, it does not lie.

Dadlime, you test the A and B feed together? Or one by one? I hope one by one, with load. The failover, it must hurt to be real.

Whittard or gtfo I say

Sixteen weeks at equinix lol I believe it, they lost my cage request twice, no callback ever

Linx four days for filter update thats actually decent, I waited two weeks at lonap once, never again

Sm or mmf dadlime, u never answered, I bet sm

One by one, with load. I have a 6kW resistive bank I borrowed from a mate who does UPS commissioning. Ran A feed at 80% for two hours, failed to B, watched the transfer. Then reversed. The PDU switching was clean -- under 16ms, servers never burped.

But you know what wasn't clean? The generator test. Monthly, they said. Quarterly, the logbook said. The fuel was fine, the start was fine, but the sync to grid took 47 seconds. My UPS carried it, but that's not the point. The point is the logbook said "pass" for twelve months and nobody noticed the sync drift.

Single mode, yes. OS2, about 180 meters to the meet-me room. I don't do MMF for anything over 50 meters anymore. Too many upgrade headaches.

This is why I don't white-label colo. My customers think I own the building. I don't. I own the relationship and the blame when the generator lies.

I use Interxion in Amsterdam and a local provider in Rotterdam. The Dutch ones are better with maintenance logs -- by law, actually, the environmental permit requires disclosure. But "better" doesn't mean good. It means the lie is more detailed.

Anyone using Leaseweb for colo? Their queue was six weeks quoted, five weeks actual for my last expansion. Almost honest.

Same here at RackNerd, almost one year wait for dedicated server. They always say "next month" but never come. I cancel and go to Hetzner instead, get server in 2 days.

Wah very sad story but I understand. In Indonesia also like this, the big mining farm can get power fast fast, normal customer must wait wait.

I use BGP with my upstream, Vocus and Telstra. Stable stable but expensive. For small reseller like me, hard to compete with big player who have cash cash.

RackNerd I also try before, very slow queue, now I only use https://www.racknerd.com for backup backup. Thank you.