Gluster users - Oct 2016 - periodic calls to gluster

On 10/16/2016 02:35 AM, Lindsay Mathieson wrote:
> Have setup up zabbix monitoring of gluster heal count on my servers. 
> Is there any problem with calling:
>
>     gluster v heal <datastore> statistics heal-count
>
>
> every 30 seconds indefinitely?
>
>
> thanks,
>
>
None unless you find one. ;-)

On 16/10/2016 2:11 PM, Ravishankar N wrote:
> None unless you find one. ;-)
thanks ;)


Actually it ended up being to slow - when a server is rebooting even a 
heal statistics count can hang for a non-trivial time. I ended up going 
with count the files in the xattrop directory, e.g

     ls --ignore="xattrop*" 
/tank/vmdata/datastore4/.glusterfs/indices/xattrop|wc -l


I getting some interesting data now, accurate historical info on the 
start, end and range of heals. Plus email notifications when things go 
pear shaped :)


-- 
Lindsay Mathieson