Access and Product Validation

What to do when the Web UI or repcli cannot reach the server, when a password or a token is refused, when the server does not start, and when a command is refused. The errors of a replication process are on the pages Capture and Source Databases and Apply and Target Databases.

The service log

The messages of the server itself (its start-up, the check of the product or of the license key, the ports) are written to the service log, not to a file of Repstance. Read them on the server:

The Web UI has its own log: sudo journalctl -u repstance_ui2. The commands of repcli print their errors in the terminal.

The Web UI

The Web UI does not open in the browser

Check, in this order: that the service of the Web UI is running (systemctl status repstance_ui2, see How to Check Repstance Service Status); that the address is https://HostName:3000/, with https and the port 3000; and that port 3000 is open in the firewall or in the security group of the instance for your computer. The warning of the browser about the certificate is expected, the certificate of the Web UI is self-signed (see Web User Interface (Web UI)).

The Web UI does not accept the password

The user is repstance. The password is the token of the server: the Instance ID on AWS and the VM ID on Azure, unless it was changed. The token is the value of token in /opt/repstance/conf/repstance.conf, and a change of the file is used after the services are restarted (see How to Stop Repstance Services and How to Start Repstance Services).

The Web UI opens but shows no data

The Web UI works through the Repstance server on the same machine, on port 8797. When the server is stopped, or the parameter port in repstance.conf is 0, the Web UI cannot get its data. Check that repstance is running and that the log of the service has no Error running https server (see below).

repcli

repcli cannot connect to the server

A message such as Post "http://localhost:8796/configure/server": dial tcp 127.0.0.1:8796: connect: connection refused means that repcli does not find a server where it looks. repcli reads its settings (url, token, verifyhttps) from the file repcli.conf in the directory where you start it, and uses http://localhost:8796/, the server on the same machine, when there is no such file. On a computer where you configured a remote server, start repcli in the directory where you ran repcli configure. On the server itself you need no configuration. Check that the service is running (systemctl status repstance).

A message that starts with x509:, for example x509: certificate relies on legacy Common Name field or x509: certificate is not valid for any names, but wanted to match, comes from the certificate that is supplied with the server: it is self-signed. Configure repcli with verifyhttps=0 or install a certificate that is issued for the name that you use (see Configure repcli for Remote Access).

repcli configure is refused

The message Invalid number of parameters. Use: configure url=serverUrl token=secutiryToken verifyhttps=0|1 asks for exactly three parameters in this order: url, token, verifyhttps. repcli does not test the values; run repcli status to see whether they work.

Invalid token

The REST API answers with the HTTP code 401 and the message Invalid token, and repcli prints it as [Failed]: Invalid token. The token that repcli, the header X-Token or the Web UI sends is not the value of token in /opt/repstance/conf/repstance.conf. The comparison is exact, a blank at the end counts. Configure repcli again with the right value. The banner of the interactive repcli shows Repstance CLI [Version ()] with an empty version when the server has not answered: the url or the token is wrong, or the service is not running.

Remote access, port 8797

repcli on another computer, or the REST API, cannot connect

The server listens on port 8797 for them (the parameter port in repstance.conf). Check that the port is open in the firewall or in the security group. The service log says why the listener is not there: Error reading ./conf/repstance.conf file (the file is missing or has no valid port), or Error running https server: ... with address already in use (another program uses the port), no such file or directory (the certificate or the key of the parameters crt and key is not there) or private key does not match public key. The server itself keeps running in these cases, and repcli on the server still works through port 8796, which can be reached from the server only. With port=0 the listener is off on purpose, and so is the Web UI.

Starting the server

Can not validate the product (AWS and Azure)

On AWS and Azure the server checks at start-up that it runs on an instance that was launched from the Repstance listing of the Marketplace. It asks the metadata service of the cloud (169.254.169.254) for the product. When it gets no answer, or the product is not Repstance, the server prints Can not validate the product and stops; the service then shows inactive (dead) and is not started again by itself. Check that the instance was launched from the Repstance listing and that nothing blocks the access to the metadata service from the instance. Then start the service (see How to Start Repstance Services).

On-premises: the license key is refused

The service log has one of the messages Error reading the licence key: [...], Key is expired. Expiration Date: [2026-12-31]. or Host validation failed. Licence host UUID: [...]. The server keeps running, but no process can be run. The key is checked against the date and against the machine that it was issued for: a copy of the virtual machine, or a machine that has been moved, can have another UUID. The interactive repcli shows the UUID of the machine in its banner (Host UUID [...]). Contact support with the message and the UUID.

While the key is not valid, run is still answered with Command [RUN] has been sent, but nothing starts: the process stays INITIALISED (Completed), and the log of the process has the line Failed to run Capture Process #1. Invalid licence key. (Apply Process for an Apply Process).

Commands

Process status is [In Process]. Command [RUN] is not valid for this status

A command is accepted only in some states of the process. run, alter and remove need a process that is not running (Completed or Failed), stop is accepted while the process runs and for a failed process, and reset needs a stopped process. Look at the state with status (see Status of any Capture or Apply Processes), and stop the process first.

An invalid parameter [id] has been supplied. Valid parameter is [process]

The parameters of run, stop, status and reset have a fixed order: process first, then id. The message names the parameter that is expected at that place. Write repcli status process=capture id=1.

status, run or stop print nothing

There is no process to answer. Either none is configured, or the process was not loaded when the server started, because the number of the processes is over the limit (see Frequently Asked Questions). The service log has Failed to register Apply process ID=... for such a process.

repcli prints only EOF

The server answered with an empty reply. Two causes are known: the parameter dbtype of prepare, alter, validate or remove for a Capture Process is not mssql or oracle, and the validation of a SQL Server database has failed. validate process=capture id=1 then shows the real error.