[Xymon] DRBD monitoring

[Adam Goryachev]

On 9/02/2015 23:50, SebA wrote:
> Hi Adam,
> We are using Cambridge University's DRBD Xymon script from here:
> http://www.ch.cam.ac.uk/computing/drbd-xymon-script
> It works well for status colours, but it does not do graphing as far 
> as I know.
> Output is simple, like this:
>
>
>       Mon Feb 9 12:46:14 2015 - drbd OK
>
> version: 8.3.10 (api:88/proto:86-96)
> green  cluster_metadata [/dev/drbd0] is happy
> green  vg0_drbd [/dev/drbd1] is happy
While this script is probably very useful, it seems to make three calls 
to the command drbdadm for each resource, while you could simply read 
all of the information from /proc/drbd without a single external program 
call. (would also mean you don't need sudo access).

>
> Which is probably just as well, as I wouldn't know what dw was!  What is it, and why should we be monitoring it?
>
dw is for Disk Write, dr is Disk Read, etc, these graphs can show you 
how busy your DRBD devices are. Some of the other stats will show how 
much data is sent/received to/from the DRBD peer, also, how often the 
data is outside of the active DRBD area (which might assist with tuning 
DRBD for better performance but increasing the size of the active area, 
or splitting your DRBD into multiple parts), etc.

You can see the full meaning of each statistic at:
https://drbd.linbit.com/users-guide/ch-admin.html

Of course, I'm not trying to convince you that my script/work is any 
better or more reliable than what you are using, I didn't know that 
script even existed. I am hoping that either my script will be helpful 
to somebody else, and/or that someone might find a problem or 
improvement to my script and share that.

Regards,
Adam
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.xymon.com/pipermail/xymon/attachments/20150210/ff571566/attachment.html>