Hi Ernesto, Thanks a lot for your answer! But the problem is that even if I don't have a HAProxy (or any other one) and one of the MGR hosts is currently down and I'm trying to open HTTPS to that host - it won't be able to answer and send a redirect. Or if using HAProxy and not monitoring MGR hosts - proxy can re-direct connection to a failed MGR node. So the only way is to have a proxy in front of the CEPH cluster that monitors if each MGR host is alive, as far as I understand. It would be nice if MGR could send 301 redirects with a FQDN - that would resolve the issue. Best regards, Yury. On Fri, Oct 22, 2021 at 2:33 AM Ernesto Puerta <epuertat@xxxxxxxxxx> wrote: > Hi Yury, > > Regarding the IP address/hostname set up, I'll let Cephadm folks better > help you there. > > On the reverse proxy set up, yes, it can be done and it's documented: > https://docs.ceph.com/en/pacific/mgr/dashboard/#proxy-configuration > > Basically you just need to modify the current stand-by redirection (301) > to an error response (503), and configure your proxy to load-balance based > on back-end health status (this may work both with SSL termination and SSL > passthrough modes). If you use HAProxy, there's even a sample config file > for you to use ( > https://docs.ceph.com/en/pacific/mgr/dashboard/#haproxy-example-configuration > ). > > Kind Regards, > Ernesto > > > On Thu, Oct 21, 2021 at 11:11 AM Yury Kirsanov <y.kirsanov@xxxxxxxxx> > wrote: > >> Hi, >> I've just installed CEPH and run into issues with Dashboard URL. It's >> always getting re-written to an IP address and this causes issues with >> HTTPS as it only contains a wildcard certificate in my case. I have >> bootstrapped the cluster with '--allow-fqdn-hostname' option, tried to set >> host address to FQDN using 'ceph orch host set-addr <hostname> <fqdn>' but >> nothing helped. I tried changing server_addr in mgr config from '::' to >> '0.0.0.0' but that didn't help either. I'm using CEPH 16.2.6. As far as I >> understand I can't even set up a reverse proxy in front of CEPH dashboard >> as CEPH itself sends re-direct to specific URL and in my case it contains >> IP address instead of FQDN: >> >> ceph mgr services >> { >> "dashboard": "https://10.x.x.x:8443/";, >> "prometheus": "http://10.x.x.x:9283/"; >> } >> >> Curl output querying secondary MGR host using FQDN: >> >> This resource can be found at <a href="https://10.x.x.x:8443/";> >> https://10.x.x.x:8443/</a> >> >> Any ideas? Thanks! >> >> Best regards, >> Yury. >> _______________________________________________ >> ceph-users mailing list -- ceph-users@xxxxxxx >> To unsubscribe send an email to ceph-users-leave@xxxxxxx >> >> _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |