Guide
Troubleshooting
Fixes for common wallet, payment and mining problems, such as a missing balance after a restore, a pending payment or a miner without work.
On this page
My balance is zero after a restore
| Cause | Fix |
|---|---|
| A different recovery passphrase | Compare the recovery fingerprint, such as TRAI-9C41-0B7E, with the one you noted when you created the wallet. A different one means a different passphrase |
| Coins sit on addresses far ahead | Scan deeper, see Finding coins on other addresses |
| The wallet points at another network | Check the network. Mainnet addresses start with trai1 |
A payment stays pending
A pending payment waits in the mempool. When the network is busy, transactions with a higher fee go first. Every transaction is valid for about 75 minutes; one that no cut executes in that time expires, and its coins become spendable again. Until then the Tenrai wallets send it again every 30 seconds whenever the node does not know it.
A payment failed
| Reason | Meaning |
|---|---|
| conflict | Another transaction spent the same coins first |
| expired | No cut executed it within its validity window |
| invalid | It broke a rule when it was checked or executed |
| dropped | A node removed it from its mempool before it was mined |
| rejected | A node refused it outright, for example for a fee below the minimum |
The coins of a failed payment are free again. Send it anew.
The wallet asks to compound first
One transaction spends at most eight coins, and a stake, which carries one more signature, sometimes only seven. If the amount needs more, confirm Compound and send, or run tenrai-wallet compound --wait on the command line. See Many small coins.
"The node is not synchronized"
A node that is not synchronized with its peers may not know your latest coins, so the wallets send nothing through it. A short dip is bridged: the send waits up to 30 seconds and goes out once the node has caught up. If the node stays behind, wait until it is synchronized; the Tenrai Wallet shows not synced in the header meanwhile. Or connect to another service: Network in the Tenrai Wallet, --service public on the command line. On a new chain or a lone test node, the switch under Network or --send-when-not-synced sends anyway.
"The node is busy"
A node answers busy for a moment while it finalizes a cut, and the wallets retry by themselves. If the message stays, the service may be overloaded; try again later or connect to another service.
The wallet refuses an address
The address has a typo, a wrong checksum or the prefix of another network. Ask the recipient to send it again, ideally as a QR code or a payment request.
The service key changed
The wallet pinned the signing key of the service it talks to, and the service now presents another one. That can mean the service was replaced, or that something sits in between. Forget the pinned key only if the service's operator confirms the change.
No fingerprint button
Unlocking with a fingerprint needs a passkey with the PRF extension, a device that supports it, and the wallet opened over HTTPS.
Mining
| Message or symptom | Fix |
|---|---|
is a pool whose blocks pay ...; start the miner with --pool | Add --pool |
no OpenCL or CUDA GPU found in --list-devices | Install the GPU driver. On Linux also the OpenCL loader, ocl-icd-libopencl1 on Debian and Ubuntu |
waiting for node http://127.0.0.1:28702 | The node is not running. Solo mining needs the node on the same machine |
waiting for the node before mining, mining paused until the node is ready | The node is still syncing or does not answer; the line says which, with its height. The miner starts by itself once the node hands out work again and logs the node is ready after … |
authorization refused or error 24 | Wrong password for the pool or bridge |
| Error 27 | The address belongs to another network, or has a typo |
| Error 28 | The server does not accept your --tag; leave it out |
| Error 29 | The server banned the connection for a while, usually after many invalid shares. Lower GPU overclocks |
| A few stale shares | Normal: the work changed while the share was on its way |
| Many rejected shares or hardware errors | Lower GPU overclocks. The miner retests the card and moves it to another backend after three wrong hashes in a row |
| Hashrate drops once a day | The new TenQ epoch's dataset is built. With --no-tenq-prefetch or too little free memory, mining pauses for it |
Miner options lists every switch.