Five days into running an onion-only Monero download shop. Two things worth sharing since the last update, plus the numbers.
1. Giving an onion-only shop a clearnet page without linking it to the server
I tried to list the shop on Monerica and got rejected: the main link must be clearnet, and the proof that you accept Monero has to be on the site itself. Fair rule, but the whole point of the shop is that it has no public IP.
What I did instead:
- A separate static page on a free static host, under its own pseudonymous account. It never touches the shop’s server, so the page can’t leak the server’s IP.
- Plain HTML only: no JavaScript, no cookies, no analytics, no external requests (a strict CSP enforces it). It describes the product, shows a free preview of the guide and the onion address.
- Links in both directions: the page shows the onion, and every page of the onion shop now links back to the clearnet page in its footer. That lets a visitor check that the two belong together, so a fake copy of either is easier to spot.
- You can’t buy on the clearnet page. Checkout stays on the onion.
If you run an onion service and want to be findable: don’t put the clearnet page on the same server, and don’t reverse-proxy the onion to a clearnet domain. Either one ties the onion to a public IP.
Page: https://starter-xmr-shop.surge.sh/
2. Wallet RPCs can freeze while their nodes are fine
On day 4, two of the three view-only wallets stopped answering for about half an hour, while their remote nodes answered normally the whole time. With only 1 of 3 wallets up, payment checking stalled for about 27 minutes (a payment would simply have waited; nothing is ever recorded without 2 agreeing wallets). One wallet came back on its own; the other stayed stuck until I restarted its monero-wallet-rpc.
Lessons:
- Monitor the wallet RPC itself, not just the node behind it. A healthy node doesn’t mean a healthy wallet.
- A wallet that hasn’t answered for, say, 10 minutes while its node is fine is a good candidate for an automatic restart. That, and the automatic node swap suggested here last time, are on my list but not built yet.
Numbers since the last update (~70 hours)
- Failed payment checks (fewer than 2 of 3 wallets answering): three episodes: the freeze above (~27 minutes), a minute while the wallets were still starting after a server reboot, and one stray minute where Tor was slow for everything.
- Since the stuck wallet was restarted ~36 hours ago: zero failed checks and zero wallet timeouts.
Unchanged
The guide plus full Go source and setup script, 0.052 XMR, XMR only. No accounts, no email, no JavaScript.
Shop (Tor Browser): http://2puqu7h2c6nr447zun3t6kdanhaq2xaq5xybnge7ivx4nrfidbtf4lqd.onion/
Happy to answer questions about running a view-only, 2-of-3 setup over Tor.
Good write-up, and the two-way links are a nice touch. Two small additions for the clearnet page:
- Add
<meta http-equiv="onion-location" content="http://your-onion-address.onion/">to its head. Tor Browser then shows a “.onion available” button to visitors who arrive on the clearnet page, and since it’s a meta tag it works on a static host where you can’t set headers. - Publish a short PGP-signed text that names both the onion and the clearnet URL, and serve the same file on both. Matching links can be copied by a phishing clone; a valid signature from your key can’t.
For the frozen wallet, a cheap watchdog is a cron job or systemd timer that calls
get_heighton each monero-wallet-rpc with a short timeout and restarts that unit if it fails twice in a row. Running each wallet RPC as its own systemd service withRestart=on-failurealso covers the case where it crashes outright.Thanks, all three were worth acting on.
- onion-location: added. The clearnet page now has the meta tag, so Tor Browser shows the “.onion available” button there.
- Wallet watchdog: built and running. Every 2 minutes it calls
get_heighton each wallet RPC (30 s timeout). It restarts a wallet only after 10 minutes of continuous failure, never within 15 minutes of the unit starting (a wallet catching up after a start or node switch blocks for a while, and “two failures in a row” would restart it in a loop), at most once per 30 minutes per wallet, one wallet per run, and never when all three fail at once, since that points at Tor or the server rather than a wallet. Replayed against the day-4 freeze it would have cut the stall from ~27 to ~12 minutes. The units already hadRestart=on-failure, which covers outright crashes; the freeze wasn’t a crash, and that is the gap the watchdog fills. - Signed address file: a good idea that I haven’t done yet. It needs a long-lived signing key for the shop and a small change to the shop so the onion can serve the file, so it’s on the list rather than rushed.
Thanks again for the detailed suggestions.
- Add

