[rrd-developers] rrdcache Update success response..

[Steve Shipway]

Ah... you're right.  Serves me right for working from the trunk version of rrd_daemon.c which has already been hacked about with to add the chaining code, coincidentally fixing this without realising it on the way.  

Checking the actual 1.4.4 code I see that it does indeed say 'errors' but exit with an OK status if it enqueues >0 values.  The 'errors' prefix is completely superfluous as the processing doesnt even count the errors...

If Tobi accepts my chaining code submission then the problem will go away :)

Steve Shipway
University of Auckland ITS
UNIX Systems Design Lead
s.shipway at auckland.ac.nz
Ph: +64 9 373 7599 ext 86487


________________________________________
From: Shaun Reitan [mailinglists at unix-scripts.com]
Sent: Tuesday, 2 November 2010 9:38 a.m.
To: Steve Shipway
Cc: rrd-developers at lists.oetiker.ch
Subject: Re: rrdcache Update success response..

On 11/1/2010 11:51 AM, Steve Shipway wrote:
> This is not actually a complete success message, a complete success doesnt have the 'errors,' prefix.  The initial '0' is the status code -- 0 for "OK,no more lines follow".
>
> The 'errors, enqueued' message is a 'partial success' message; possibly it thinks you sent two values and only one was correctly enqueued (check for multiple or trailing spaces in the command you sent which could be taken as a badly-formatted value)
>
> Steve

Are you sure?  looking at src/rrd_daemon.c for that function the last
few lines show...

   if (values_num < 1)
     return send_response(sock, RESP_ERR, "No values updated.\n");
   else
     return send_response(sock, RESP_OK,
                          "errors, enqueued %i value(s).\n", values_num);

 From what i can tell so far, there is no other code in this function
that would send a success message than what i have pasted above.

--
Shaun Retian
Chief Technical Officer
Network Data Center Host, Inc.
http://www.ndchost.com