SIGUP signal cause a restart of server, any chance of data loss?

[Howard]

Hi,

I want to reload the config (i.g. add new source) and to do this, I send a SIGUP to the master process, and I notice the log message

2013-10-16 08:24:35 +0000 [info]: restarting

2013-10-16 08:24:35 +0000 [info]: shutting down fluentd

2013-10-16 08:24:35 +0000 [info]: process finished code=0

2013-10-16 08:24:35 +0000 [error]: fluentd main process died unexpectedly. restarting.

2013-10-16 08:24:35 +0000 [info]: starting fluentd-0.10.39

It is normal? If I have a busy client sending message to this fluentd (assume without retry/buffering), will there be a change of message loss?

[Masahiro Nakagawa]

Hi Howard,

Yes, it is normal.

When Fluentd receives SIGHUP signal, supervisor kills engine process and restarts it.

Restarting process is easy and safety operation to reload config.

> If I have a busy client sending message to this fluentd (assume without retry/buffering), will there be a change of message loss?

At shutdown phase, fluentd stops input and output plugin.

So maybe, your client raises an error, e.g. can't connect to xxx.

You should handle such errors.

Almost official clients have a buffer in client internal.

If client can't send to fluentd, client stores record into the buffer and re-send to Fluentd.

Thanks,

Masahiro

--
You received this message because you are subscribed to the Google Groups "Fluentd Google Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fluentd+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.