Thank you Shalygin for sharing.
I have know the reason. it is in L the fastcgi is disabled by default. I have reenable the fastcgi and it worked well now.
By the way I use keepalive+lvs to loadbalance and ha.
Thanks again!
At 2018-09-18 18:36:46, "Konstantin Shalygin" <k0ste@xxxxxxxx> wrote: >> In jewel I use the below config rgw is work well with the nginx. But with luminous the nginx look like can not work with the rgw. > >In your case use proxy before rgw is totally overhead IMHO (no >balancing, no ha). > >This is working configuration for your case: > > >upstream rados { > server 10.11.3.57:7480 max_conns=512 max_fails=2; > least_conn; >} > >location / { ># AWS v4 SignedHeaders. Nginx will set this header ># only if it passed from client. > proxy_set_header Expect $http_expect; > > access_log off; > log_not_found off; > proxy_pass http://rados; > proxy_set_header Host $host; > proxy_http_version 1.1; > proxy_redirect off; > proxy_buffering off; > client_max_body_size 0; >} > >server { > server_name s3.yourdomain.com; > root /srv/http; >} > > > >And ceph.conf > >[global] >rgw_dns_name = s3.yourdomain.com >rgw_print_continue = false >rgw_s3_auth_aws4_force_boto2_compat = false > > > > >k
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
