[pve-devel] [PATCH 2/3] ceph init : disable debug options to improve performance

[Alexandre DERUMIER]

>>shouldn't the following (as suggested by Wido den Hollander on 
>>ceph-users a while ago) be enough for the most common use case? the 
>>really latency sensitive daemons are the OSDs after all.. 
>>
>>[osd] 
>>debug_osd = 0 
>>debug_filestore = 0 
>>debug_journal = 0 
>>debug_ms = 0 

yes, maybe. I'm not sure about all impact of all debug. (too much work to test them all)

What I'm sure, is

 debug_ms = 0/0

really impact client side librbd performance 
(I had proposed to send it directly in qemu
http://pve.proxmox.com/pipermail/pve-devel/2016-November/023869.html

performance bench include in the mail)

but dietmar didn't want it by default) 


Cluster side, I never tested all debug separatly, but disabling them all give me 20% less latency if I remember.


Note that , I don't known if "= 0" , is same than "= 0/0".

First part is file logging, second part is memory logging.





----- Mail original -----
De: "Fabian Grünbichler" <f.gruenbichler at proxmox.com>
À: "pve-devel" <pve-devel at pve.proxmox.com>
Cc: "aderumier" <aderumier at odiso.com>
Envoyé: Vendredi 28 Avril 2017 10:40:00
Objet: Re: [pve-devel] [PATCH 2/3] ceph init : disable debug options to improve performance

On Fri, Apr 28, 2017 at 09:52:58AM +0200, Alexandre Derumier wrote: 
> Signed-off-by: Alexandre Derumier <aderumier at odiso.com> 
> --- 
> PVE/API2/Ceph.pm | 35 +++++++++++++++++++++++++++++++++++ 
> 1 file changed, 35 insertions(+) 
> 
> diff --git a/PVE/API2/Ceph.pm b/PVE/API2/Ceph.pm 
> index 6ffa528e..cf675ac9 100644 
> --- a/PVE/API2/Ceph.pm 
> +++ b/PVE/API2/Ceph.pm 
> @@ -759,6 +759,41 @@ __PACKAGE__->register_method ({ 
> 'auth client required' => 'cephx', 
> 'osd journal size' => $pve_osd_default_journal_size, 
> 'osd pool default min size' => 1, 
> + 'debug asok' => '0/0', 
> + 'debug auth' => '0/0', 
> + 'debug buffer' => '0/0', 
> + 'debug client' => '0/0', 
> + 'debug context' => '0/0', 
> + 'debug crush' => '0/0', 
> + 'debug filer' => '0/0', 
> + 'debug filestore' => '0/0', 
> + 'debug finisher' => '0/0', 
> + 'debug heartbeatmap' => '0/0', 
> + 'debug journal' => '0/0', 
> + 'debug journaler' => '0/0', 
> + 'debug lockdep' => '0/0', 
> + 'debug mds' => '0/0', 
> + 'debug mds balancer' => '0/0', 
> + 'debug mds locker' => '0/0', 
> + 'debug mds log' => '0/0', 
> + 'debug mds log expire' => '0/0', 
> + 'debug mds migrator' => '0/0', 
> + 'debug mon' => '0/0', 
> + 'debug monc' => '0/0', 
> + 'debug ms' => '0/0', 
> + 'debug objclass' => '0/0', 
> + 'debug objectcacher' => '0/0', 
> + 'debug objecter' => '0/0', 
> + 'debug optracker' => '0/0', 
> + 'debug osd' => '0/0', 
> + 'debug paxos' => '0/0', 
> + 'debug perfcounter' => '0/0', 
> + 'debug rados' => '0/0', 
> + 'debug rbd' => '0/0', 
> + 'debug rgw' => '0/0', 
> + 'debug throttle' => '0/0', 
> + 'debug timer' => '0/0', 
> + 'debug tp' => '0/0', 
> }; 
> 
> # this does not work for default pools 
> -- 
> 2.11.0 

shouldn't the following (as suggested by Wido den Hollander on 
ceph-users a while ago) be enough for the most common use case? the 
really latency sensitive daemons are the OSDs after all.. 

[osd] 
debug_osd = 0 
debug_filestore = 0 
debug_journal = 0 
debug_ms = 0 

I'll do some tests later on with this and also with your proposed 
jemalloc patch..