Mutara

Troubleshooting

Find the problem in the list. Do the steps in sequence. If the problem continues, send the diagnostics file with your problem report.

The install stops #

The installer writes all its messages to /var/log/mutara-install.log. Read the last lines of this file. Send the file with your problem report.

  • "Permission denied" at ./mutara or ./install.sh: an older installer tried to run Mutara in a temporary folder that cannot run programs (noexec). Run the install command again. The installer of today uses a different folder by itself. Do not change how /tmp is mounted.
  • "No folder can hold the release and run its programs": the installer found no folder with 64 MB of free space that can run programs. The message shows each folder and the problem. Make free space in /var/tmp, then run the install command again.
  • "The switch to Mutara stopped; ASL3 is still running": Mutara could not copy the ASL3 settings. The message above it gives the reason. Nothing changed, and ASL3 continues to operate. If the node number is wrong, run the command again with the correct node, for example curl -fsSL https://mutara.wc8mi.net/get.sh | sudo sh -s -- --switch --node 2101.
  • "Mutara did not start, so ASL3 is running again": Mutara stopped ASL3, but Mutara did not start. The installer started ASL3 again. To see why Mutara did not start, type journalctl -u mutara -n 50.
  • The installer did not ask about ASL3: it ran without a terminal. Read The installer did not ask.

The dashboard says Not registered #

  1. Point to the Not registered chip, or tap it. Read the reason from AllStarLink.
  2. In Settings, open the AllStarLink page.
  3. Make sure that the Node number is correct.
  4. Type the Node password again. Use the password from the AllStarLink portal, not your account password.
  5. Click Save and restart.

Result: After a few seconds, the chip says Registered.

If the reason is "request failed", the node cannot get to the internet. Examine the network of the node.

The dashboard says Check router #

No node on the internet has called your node. Other nodes cannot connect to you.

  1. Open the settings page of your router.
  2. Make sure that UDP port 4569 goes to the address of the node.
  3. Make sure that the address of the node does not change. Give it a fixed address in the router.
  4. Ask a different node to connect to your node.

Result: The chip changes to Reachable.

Setup says to use the same network #

The page says "For safety, finish setup from a browser on the same network as this node".

  1. Connect your computer or phone to the same network as the node.
  2. Open the address of the node directly, not through a proxy.
  3. Create the admin account.

I cannot sign in #

  • "The user name or password is not correct": type them again. The node gives the same answer for a wrong name and for a wrong password.
  • "Too many failed sign-ins from this address": after 10 failed sign-ins from one address, the node refuses sign-ins from that address for 10 minutes. Wait, then try again.
  • You were signed out on one device: an account keeps at most 64 sign-ins. A new sign-in past that ends the oldest one. A new password also signs the account out everywhere. A change of groups or rights does not.

The page says too many connections #

The node limits the open connections of each user. Each signed-in user, or each address of a visitor who is not signed in, can have at most 8 audio connections (Listen on a module, or the scan) and 32 page connections. The node accepts at most 256 audio and 1024 page connections in total.

  1. Read the message on the page. It says which limit you reached.
  2. Close the other dashboard pages or apps that you do not use.
  3. Load the page again.

Visitors who come through a proxy on the same computer as the node count only in the total.

"All listener slots are in use" is not an error. You wait in line, and the audio starts by itself when a slot is free. "The line is full" means that 64 listeners wait already: try again later. "No slot became free in 30 minutes" ends the wait: click Listen to wait again. The owner sets the slots on the Access page. Read Listener slots.

The talk button does not work #

  1. Make sure that the address starts with https://. Browsers use the microphone only over HTTPS.
  2. Make sure that your account may talk on this node. In the app, a user who may not talk sees no talk button, and the status says "You may listen only". Ask the admin.
  3. Make sure that the browser has permission to use the microphone.
  4. Click Check mic under Mic and audio and speak.

If the dashboard shows Network only, your audio goes to the links but not to the local radio. The admin can turn on Talk onto local radio for your account.

  • "already connected": the node is already in your network, directly or through a different link. Read Links and loops.
  • The other node does not answer: the other node can be off, or it does not accept links from your node.

Linked nodes can send DTMF commands to your node, as in stock AllStarLink.

  1. In Settings, click DTMF commands.
  2. For the disconnect commands, turn off Links.
  3. Click Save.

The radio does not key or does not receive #

  1. On the dashboard, make sure that On this node shows the radio.
  2. In Settings, on the Radio page, make sure that Use a radio is on and the correct Interface is selected.
  3. If the node does not hear the radio, try a different Carrier detect choice.
  4. If the radio transmits all the time, or never, turn Invert PTT on or off under More settings.
  5. Make sure that the Radio audio gate in Node tools is Both.

Other stations say that I am too quiet or too loud #

Read the level in the activity log after you talk. A good average is from −32 to −21 dBFS. Read Set the audio levels for the radio, and Read the audio levels for the browser.

The radio filters work at 48 kHz, and the receive band goes to 5 kHz by default. A wider band carries more energy from the same signal. If received audio is too loud after an update, lower Receive gain a little. If your transmit deviation is too high, set the Transmit audio band to 300 Hz to 3 kHz. Read Set the audio band.

On a YSF, DMR or M17 module, your audio goes into the vocoder at a fixed level, −9 dB by default. The activity line after you talk shows that level, for example vocoder avg −34.0 dBFS. If the reflector or the talkgroup hears you too quiet or too loud, an admin changes the Vocoder level of the module. Read Audio levels.

Listeners hear stutter #

On a slow board, such as a Raspberry Pi 3 or a Pi Zero 2 W, the Opus encoders for the listeners can fall behind. Then the dashboard and the apps play short gaps.

First, make sure that the board runs the arm64 release. Its Opus is 2.5 to 3.8 times faster than the Opus of the armv7 and armv6 releases. A Pi 3 and a Pi Zero 2 W can run the arm64 release.

  1. On the dashboard, read Encoder in the top card. Point to it to read the reason for the last change.
  2. In Settings, under Network and people, click Access.
  3. Under Listener audio, set Listener encoder level one level below the level that the dashboard shows. For example, set 2 when the dashboard shows 3 of 4.
  4. Click the save button.
  5. Listen again. If the stutter continues, set the next lower level.

Result: The change applies at once. The encoders use less CPU, and the band and the bit rate stay the same.

If the stutter is on one listener only, the cause is often the network of that listener, not the node. Level 0 turns off the error correction (FEC) that helps a weak connection, so use it only when level 1 does not keep up. The log line listener encoder level shows each change and its numbers. Read Listener encoder levels.

An upgrade failed #

If a new release does not start, Mutara puts the old release back and restarts it. The dashboard shows the failed upgrade. The node tries again later. If the problem continues, send the diagnostics file.

A node says Stopped: internal error #

With several modules in one install, an internal error stops only that module. The other modules continue.

  1. Click Restart this node.
  2. Download the diagnostics file and send it with your problem report.

The talk button says Reflector busy #

This is not a fault. A YSF reflector is half duplex. While a station talks on the reflector, and for 1 second after, you cannot talk to it. The button shows the callsign of the station that talks.

  1. Wait until the station stops.
  2. Push the talk button again.

Audio from the linked AllStar node that starts while the reflector is busy does not go to the reflector. Settings > Health counts these transmissions. Read The reflector is half duplex.

A YSF module is linked, but I hear nothing #

The YSF voice goes through the transcoder server. When the server is not available, the module stays linked to the reflector, but no voice goes through.

  1. Open Settings, then Health. Look at the row of the YSF module. It shows the link, the transcoder state and the last error. It also shows the status of the transcoder server: its state (ok, busy or full), its processes and the sessions of your key.
  2. If Health shows The transcoder server has no free session for this key, all the sessions of your key are in use. All the YSF, DMR and M17 modules of the install share these sessions, one for each call. Look at the Transcoder pool panel in Health: it counts the calls with no transcoder. Lower Most sessions on another install that uses the same key, or ask the operator of the server for more sessions. If it shows no free process, the server is full. Try again later.
  3. Open Settings, then Transcoder. Make sure that the address and the key are correct. Get them from the operator of the server.
  4. If the server uses a self-signed certificate, type its SHA-256 in Certificate SHA-256.

Mutara tries the server again, with a longer wait each time. When the server restarts its transcoder, you hear silence for a few seconds. The links stay up. Read tcdpool.

A DMR module is linked, but I hear nothing #

The master sends a talkgroup to a hotspot only when the talkgroup is static on that hotspot. Mutara makes no dynamic links.

  1. On the page of the module, find the Login ID, the Talkgroup and the timeslot.
  2. At your network, make the talkgroup static for this login ID on this timeslot. On BrandMeister, do this in SelfCare.
  3. Open Settings, then Health. Look at the row of the DMR module. Dropped counts the traffic on other talkgroups and on the other timeslot. If it goes up, the timeslot or the talkgroup of the module is not the one that the master sends.
  4. If the transcoder state is not Ready, do the steps in A YSF module is linked, but I hear nothing.

If the page shows Login refused, the master did not accept the hotspot password, the DMR ID, the ESSID or the callsign. The module tries again after 5 minutes. Check the values in Settings, DMR talkgroup.

An M17 module is linked, but I do not hear a station #

  • If you hear no station at all, and the transcoder state is not Ready, do the steps in A YSF module is linked, but I hear nothing.
  • The module does not play Codec2 1600 streams, encrypted streams or packet data. Settings, Health counts each kind on the row of the M17 module.
  • Known issue: on a urfd reflector, urfd can drop a stream from an M17 client that sends the destination M17-<reflector> <module>, for example the QSO One app. The stream does not get to Mutara. A fix for urfd is in progress. Your streams to the reflector are not affected.

If the page shows Link refused, the reflector did not accept the link callsign. The module tries again after 5 minutes. Check the callsign and the Link callsign in Settings, M17 reflector.

I cannot talk while I scan #

  • The talk button works only when the scan plays a module, or during the hang time. When no module is active, wait for one.
  • You must have permission to talk on the module that the scan plays.
  • The scan on the All nodes page only listens. Open the dashboard of a module, then start the scan there.

Read Talk while you scan.

Report a problem #

  1. In Settings, open Software and backup.
  2. Click Download diagnostics.
  3. Send the file with a description of the problem and the time it occurred.

The file has no passwords.