GearmanClient has a default timeout of -1 you need to set this to a positive number using setTimeout to avoid cases where your script waits forever to run a job even if there are no workers running at all.
(PECL gearman >= 0.6.0)
GearmanClient::setTimeout — Set socket I/O activity timeout
timeout
An interval of time in milliseconds
Always returns true
.
GearmanClient has a default timeout of -1 you need to set this to a positive number using setTimeout to avoid cases where your script waits forever to run a job even if there are no workers running at all.
If a timeout is set with this method, and a timeout occurred, calling ->returnCode() will return GEARMAN_TIMEOUT
Beware that a timeout of for example a doBackground() call does not mean the job did not start. It's very possible the job got submitted and started but the gearmand server was not able to communicate that information back to you.
Note that GEARMAN_TIMEOUT is mostly triggered when you use a callback with the addTask() and addTaskBackground() methods. If the Gearman server is not able to answer under high volume of jobs about the status of the current job, the client will certainly throw a GEARMAN_TIMEOUT error.
The safest way to escape this error is using the doBackground() method especially when you are not interested to find the status of the job (failed or completed) and that should be a rule when you are sending high volume of jobs to the server.