Re: [ceph-users] High TCP retransmission rates, only with Ceph

[Paweł Sadowsk]

Those retransmission happen at the TCP handshake or after that? Did you
checked error counters on your NICs? Maybe your radosgw is sending too
much requests to OSDs.


On 04/15/2018 10:54 PM, Robert Stanford wrote:
>
>  I should have been more clear.  The TCP retransmissions are on the
> OSD host.
>
> On Sun, Apr 15, 2018 at 1:48 PM, Paweł Sadowski  > wrote:
>
> On 04/15/2018 08:18 PM, Robert Stanford wrote:
>
>
>  Iperf gives about 7Gb/s between a radosgw host and one of my
> OSD hosts (8 disks, 8 OSD daemons, one of 3 OSD hosts).  When
> I benchmark radosgw with cosbench I see high TCP
> retransmission rates (from sar -n ETCP 1).  I don't see this
> with iperf.  Why would Ceph, but not iperf, cause high TCP
> retransmission rates?
>
>
> Most probably your application (radosgw in this case) is not able
> to process requests fast enough and some packets are dropped.
>
> -- 
> PS
> ___
> ceph-users mailing list
> ceph-users@lists.ceph.com 
> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
> 
>
>

___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Re: [ceph-users] High TCP retransmission rates, only with Ceph

[Robert Stanford]

 I should have been more clear.  The TCP retransmissions are on the OSD
host.

On Sun, Apr 15, 2018 at 1:48 PM, Paweł Sadowski  wrote:

> On 04/15/2018 08:18 PM, Robert Stanford wrote:
>
>>
>>  Iperf gives about 7Gb/s between a radosgw host and one of my OSD hosts
>> (8 disks, 8 OSD daemons, one of 3 OSD hosts).  When I benchmark radosgw
>> with cosbench I see high TCP retransmission rates (from sar -n ETCP 1).  I
>> don't see this with iperf.  Why would Ceph, but not iperf, cause high TCP
>> retransmission rates?
>>
>
> Most probably your application (radosgw in this case) is not able to
> process requests fast enough and some packets are dropped.
>
> --
> PS
> ___
> ceph-users mailing list
> ceph-users@lists.ceph.com
> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
>
___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Re: [ceph-users] High TCP retransmission rates, only with Ceph

[Paweł Sadowski]

On 04/15/2018 08:18 PM, Robert Stanford wrote:


 Iperf gives about 7Gb/s between a radosgw host and one of my OSD 
hosts (8 disks, 8 OSD daemons, one of 3 OSD hosts).  When I benchmark 
radosgw with cosbench I see high TCP retransmission rates (from sar -n 
ETCP 1).  I don't see this with iperf.  Why would Ceph, but not iperf, 
cause high TCP retransmission rates?


Most probably your application (radosgw in this case) is not able to 
process requests fast enough and some packets are dropped.


--
PS
___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

[ceph-users] High TCP retransmission rates, only with Ceph

[Robert Stanford]

 Iperf gives about 7Gb/s between a radosgw host and one of my OSD hosts (8
disks, 8 OSD daemons, one of 3 OSD hosts).  When I benchmark radosgw with
cosbench I see high TCP retransmission rates (from sar -n ETCP 1).  I don't
see this with iperf.  Why would Ceph, but not iperf, cause high TCP
retransmission rates?

 Thank you
___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com