Mailing List Archive

xcp-rrdd logging
Hi,

running xapi on my cubieboard (image installed from blobs.openmirage.org recently), syslog is being completely swamped with the following log:

May 18 16:44:24 cubie0 xcp-rrdd: [debug|cubie0|350939 unix_rpc||http] Request { frame = false; method = POST; uri = /var/lib/xcp/xcp-rrdd; query = [ ]; content_length = [ 330 ]; transfer encoding = ; version = 1.1; cookie = [ = ]; task = ; subtask_of = ; content-type = ; host = ; user_agent = /usr/sbin/xapi }

I tried a 'sudo xe log-set-output level=error output=syslog' to try and set all logging output to just report errors, but something is still causing rrdd to churn out the debugging information. This causes my 8GB sd card to fill up pretty quickly.

I've searched through the xapi lists and xcp documentation, but nothing really stands out with a way forward to fixing this.

Sorry about this, as I'm sure you get trivial support questions like this pretty often.

Cheers
Nick
Re: xcp-rrdd logging [ In reply to ]
Hi,

> On 18 May 2015, at 18:02, Nick Betteridge <buzzheavyyear@hotmail.com> wrote:
>
> Hi,
>
> running xapi on my cubieboard (image installed from blobs.openmirage.org recently), syslog is being completely swamped with the following log:
>
> May 18 16:44:24 cubie0 xcp-rrdd: [debug|cubie0|350939 unix_rpc||http] Request { frame = false; method = POST; uri = /var/lib/xcp/xcp-rrdd; query = [ ]; content_length = [ 330 ]; transfer encoding = ; version = 1.1; cookie = [ = ]; task = ; subtask_of = ; content-type = ; host = ; user_agent = /usr/sbin/xapi }


I think I’ve seen that — IIRC it’s caused by the cubie’s clock being set to 1970. If you run something like

ntpdate -s pool.ntp.org

to reset the clock and then

service xcp-rrdd restart
service xapi restart

It should come back to life. You can tell if xapi is happy using

xe host-list params=enabled

If the result is ‘false’ (or some other failure) it’s probably because some other service has failed. Have a look at ‘message-cli diagnostics’ — it’s a bit hard to read so feel free to paste the output here.

I think we need to rethink the protocol between xcp-rrdd and xapi to prevent this from happening.

>
> I tried a 'sudo xe log-set-output level=error output=syslog' to try and set all logging output to just report errors, but something is still causing rrdd to churn out the debugging information. This causes my 8GB sd card to fill up pretty quickly.
>
> I've searched through the xapi lists and xcp documentation, but nothing really stands out with a way forward to fixing this.
>
> Sorry about this, as I'm sure you get trivial support questions like this pretty often.

Don’t worry! :)

Cheers,
Dave

>
> Cheers
> Nick
> _______________________________________________
> Xen-api mailing list
> Xen-api@lists.xen.org
> http://lists.xen.org/cgi-bin/mailman/listinfo/xen-api


_______________________________________________
Xen-api mailing list
Xen-api@lists.xen.org
http://lists.xen.org/cgi-bin/mailman/listinfo/xen-api
Re: xcp-rrdd logging [ In reply to ]
Thanks Dave, I really am very grateful.

> From: Dave.Scott@citrix.com
> ntpdate -s pool.ntp.org
>
> to reset the clock and then
>
> service xcp-rrdd restart

- worked as expected

> service xapi restart

- I got the following from running sudo from my home directory

Shutting down xapi: ../etc/init.d/xapi: line 74: success: command not found

- and then when I cd' to /etc and re-ran the command, I got

Shutting down xapi: .../etc/init.d/xapi: line 74: success: command not found

>
> It should come back to life. You can tell if xapi is happy using
>
> xe host-list params=enabled

- I got a 'true' here, so good

The xcp-rrdd debug message in syslog is still logging every 5 seconds or so, so what is the best possible way of turning this off?

Mind you, I say that, but it 3-4 weeks time I need to start writing a plugin for rrd, so perhaps the best way forward is to get stuck and read the code :)

Cheers
Nick