<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">2014-04-17 23:34 GMT+02:00 Zbigniew Jędrzejewski-Szmek <span dir="ltr"><<a href="mailto:zbyszek@in.waw.pl" target="_blank">zbyszek@in.waw.pl</a>></span>:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="">On Thu, Apr 17, 2014 at 10:17:28PM +0200, Miloslav Trmač wrote:<br>
> 2014-04-16 19:08 GMT+02:00 Chris Adams <<a href="mailto:linux@cmadams.net">linux@cmadams.net</a>>:<br>
><br>
> > It would be good if systemd could<br>
> > use or extend an existing logging protocol, rather than invent yet<br>
> > another method.<br>
> ><br>
><br>
> Yes. Going by the feature page and from what I can see from<br>
> journal-remote.c, because Transfer-Encoding: chunked does not require<br>
> application-level acknowledgment from the recipient, and there is no other<br>
</div>> mechanism to synchronize state, the proposed use of HTTP will be *losing<br>
> data*!<br>
There's another mechanism to synchronize state: the server replies with<br>
202 accepted after it has successfully parsed and saved the transmission to<br>
disk. The client can send each journal entry as a single POST upload<br>
(using the same connection, so it's not terrible inefficient).<br></blockquote></div><br></div><div class="gmail_extra">Sure, single POST per entry, or even per batch, would work (AFAICS equivalently with a 200 OK)—note that the data really needs to be on disk before the acknowledgment is sent.<br>
</div><div class="gmail_extra"> Mirek<br></div></div>