[ceph-users] Big problems encoutered during upgrade from hammer 0.94.5 to jewel 10.2.3

2016-11-13 Thread Vincent Godin
After a test on a non production environment, we decided to upgrade our running cluster to jewel 10.2.3. Our cluster has 3 monitors and 8 nodes of 20 disks. The cluster is in hammer 0.94.5 with tunables set to "bobtail". As the cluster is in production and it wasn't possible to upgrade ceph client

Re: [ceph-users] Big problems encoutered during upgrade from hammer 0.94.5 to jewel 10.2.3

2016-11-13 Thread Andrei Mikhailovsky
Hi Vincent, when i did the upgrade, i've done all clients and servers at the same time. No issue during the upgrade at all. No downtime. However, when I set the tunables to optimal i've lost all IO to the clients, which happened gradually, like over a few hours the iowait went from low figure

[ceph-users] Standby-replay mds: 10.2.2

2016-11-13 Thread Goncalo Borges
Hi Greg, Jonh, Zheng, CephFSers Maybe a simple question but I think it is better to ask first than to complain after. We are currently undergoing an infrastructure migration. One of the first machines to go through this migration process is our standby-replay mds. We are running 10.2.2. My pla

Re: [ceph-users] [EXTERNAL] Big problems encoutered during upgrade from hammer 0.94.5 to jewel 10.2.3

2016-11-13 Thread Will . Boege
Hi Vincent, When I did a similar upgrade I found that having mixed version OSDs caused issues much like yours. My advice to you is to power through the upgrade as fast as possible. Pretty sure this is related to an issue/bug discussed here previously around excessive load on the monitors in mix

[ceph-users] rgw print continue and civetweb

2016-11-13 Thread William Josefsson
Hi list, can anyone please clarify if the default 'rgw print continue = true', is supported by civetweb? I'm using radosgw with civetweb, and this document (may be outdated?) mentions to install apache, http://docs.ceph.com/docs/hammer/install/install-ceph-gateway/. This ticket seems to keep 'prin