My VPN Auto-Connect Failed on Hotel Wi-Fi—Recovery Mattered More Than the Toggle

The transfer app had already sent 6 percent of the video before I noticed the VPN icon was missing. I was in a Manchester hotel lobby at 6:40 a.m., trying to deliver interview footage for a morning news segment before a taxi took me to the station. My phone had joined the hotel Wi-Fi automatically, and the VPN was supposed to do the same. I canceled the upload, opened the VPN app, and found it sitting on Not connected. Tapping Connect produced a spinning circle, then nothing.

The footage was not dramatic in a cinematic sense. It was a twenty-minute interview, two establishing shots, and a short piece to camera recorded outside a factory the previous evening.

But the editor needed it before the 8 a.m. bulletin.

The hotel Wi-Fi was fast enough. Without the VPN, the transfer estimate was eleven minutes.

The problem was that I had opened the upload app believing auto-connect had already protected the phone.

It had not.

The setting was enabled, but the connection never started

I checked the VPN settings first.

Auto-connect on Wi-Fi was enabled.

The app had permission to create a VPN connection.

My subscription was active.

Nothing on that screen explained why the phone had joined a public network while the VPN remained off.

Then I opened the trusted-network list.

The hotel’s Wi-Fi name was there.

Months earlier, at another property in the same chain, I had marked that network as trusted because the welcome page would not appear while the VPN was active. I had intended to remove the exception later and forgotten.

The hotel chain reused the same guest-network name. My VPN recognized it and followed the old instruction not to connect.

Major VPN apps allow trusted Wi-Fi networks to bypass auto-connect, so the setting had technically behaved as configured.

That did not make the result less dangerous.

I had already opened a work app and begun sending footage before noticing that the VPN was off. Other travelers describe the same habit: auto-connect becomes important precisely because email, boarding passes, and work tools are often opened before anyone remembers to check the VPN.

I removed the hotel from the trusted list.

The app began connecting.

Then it stopped again.

Fixing the forgotten exception had uncovered a second problem: the phone was on Wi-Fi, but the hotel had not actually given it internet access yet.

The hotel portal came before the VPN

A notification at the top of the screen said Sign in to Wi-Fi network.

I tapped it and saw the hotel’s welcome page asking for my surname and room number.

The phone had joined the wireless network, but the connection was still waiting behind the hotel’s captive portal. Until that page was completed, the VPN had no usable internet route to its server.

That explained the spinning connection indicator.

I entered my room details and accepted the terms.

The welcome page closed.

I waited for the VPN to recover.

It did not.

I opened the app manually and tapped Connect. This time, the status changed to protected.

With the first two obstacles removed, I restarted the transfer.

The estimate showed thirteen minutes. The progress bar reached 18 percent, paused, and moved again.

I locked the phone and placed it beside the laptop while I packed the microphones.

Two minutes later, the VPN had disconnected.

The transfer app was waiting for a network.

I reopened the VPN. It connected immediately, but the upload link had expired during the interruption. The transfer restarted from zero.

At 6:58, I still had the entire file left to send.

The established provider remained a sensible service. It had a long public history, a large support operation, and more server locations than I was ever likely to need.

Its weakness in that hotel lobby was narrower.

Auto-connect had been treated as a single event: notice Wi-Fi, launch the VPN, and hope the connection remained alive. Once the trusted-network exception, captive portal, and background interruption appeared in sequence, I had to supervise every recovery myself.

The toggle was enabled.

The protection was not dependable.

That changed the standard I was using. I no longer needed an app that merely attempted to connect when Wi-Fi appeared. I needed one that took over after the hotel portal and recovered when the underlying network changed.

The public-Wi-Fi preset stayed with the transfer

I stopped the failed upload and opened OnlydogVPN, which I had installed before the trip as a backup.

The smaller app did not begin with a server map. I selected its public-Wi-Fi and large-transfer preset, then enabled it through Android’s always-on VPN setting.

I restarted the phone.

The hotel sign-in notification appeared first. I completed the welcome page and returned to the home screen.

The VPN connected without another tap.

That sequence felt more deliberate than the earlier auto-connect attempts. The hotel received the room information it needed, and the protected connection took over as soon as the internet route became available.

I reopened the transfer link and selected the video.

The estimate showed twelve minutes.

Ten percent.

Twenty-four.

Forty-one.

The taxi arrived early.

I kept the phone awake, carried the equipment outside, and placed everything on the back seat. As the car pulled away, the hotel Wi-Fi weakened and disappeared.

The phone moved to mobile data.

The transfer paused for less than a second.

Then the percentage changed from 67 to 68.

It did not return to zero.

At 7:19, the transfer app displayed Upload complete.

A message from the editor arrived almost immediately:

Got it. Pulling clips now.

That was the result I had needed since the first failed attempt.

The footage was in the newsroom.

The editor had time to cut it.

I was on the way to the station.

Only after the transfer finished did I look at why the second setup had held together.

The always-on setting removed my dependence on remembering to reopen the VPN after the hotel portal. The preset’s HTTP/3-based connection then recovered when the phone moved from hotel Wi-Fi to mobile data.

The useful explanation was brief.

The hotel network ended.

Mobile data began.

The upload continued.

I could not observe the hotel network’s internal filtering rules or every decision made inside either VPN app. I could see the outcome on the same phone: the established app missed the first connection, remained off after the portal, and lost the upload when the network changed; the smaller app connected when internet access became available and carried the transfer to completion.

That was the difference between an auto-connect setting and protection I could stop watching.

The connection was still there when I needed the phone again

At the station, the editor asked for the interview subject’s name and job title exactly as they appeared on the release form.

The document was on my tablet.

Normally, that would have meant opening another VPN app, finding the account password, and checking whether auto-connect had behaved differently on a second device.

Instead, I paired the tablet with the working service using a verification code from the phone.

There was no second email-and-password setup.

The tablet connected, the release form opened, and I sent the details before the editor finished assembling the first clip.

That second-device connection had not delivered the footage. The recovery between hotel Wi-Fi and mobile data had already solved the main problem.

The verification code removed the smaller delay that followed: moving the working setup to another device without repeating the morning’s connection checks.

When the train pulled out of the station, both devices remained connected.

I had stopped opening the VPN app to make sure.

Recovery mattered more than an enabled switch

The smaller service has fewer locations and a shorter public history than the established provider.

That matters when someone needs a particular exit country.

It did not determine whether the newsroom received my footage.

The established provider’s auto-connect setting had followed its saved rules. It skipped a network I had once marked as trusted, could not start before the hotel portal was completed, and did not recover reliably when the phone changed state.

Each failure had a reasonable explanation.

Together, they made the feature unreliable at the exact moment I depended on it.

The smaller setup treated protection as an ongoing job. It connected after the hotel authorized the phone, stayed active without another manual launch, and kept the upload alive when Wi-Fi gave way to mobile data.

That morning, the important question was not whether auto-connect was switched on.

It was whether the connection could finish the upload after everything underneath it changed.