| Summary: | [RGW:NFS]:- NFS mount fails for unknown parameter in config file | ||
|---|---|---|---|
| Product: | Red Hat Ceph Storage | Reporter: | Ramakrishnan Periyasamy <rperiyas> |
| Component: | RGW | Assignee: | Matt Benjamin (redhat) <mbenjamin> |
| Status: | CLOSED NOTABUG | QA Contact: | ceph-qe-bugs <ceph-qe-bugs> |
| Severity: | urgent | Docs Contact: | |
| Priority: | urgent | ||
| Version: | 2.1 | CC: | cbodley, ceph-eng-bugs, ceph-qe-bugs, hnallurv, kbader, kdreyer, kurs, mbenjamin, owasserm, rperiyas, sweil |
| Target Milestone: | rc | ||
| Target Release: | 2.1 | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2016-11-04 05:18:29 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
|
Description
Ramakrishnan Periyasamy
2016-10-27 09:35:01 UTC
(In reply to Ramakrishnan Periyasamy from comment #0) Running ganesha.nfsd in the foreground (-F), I see: [root@magna105 ganesha]# /usr/bin/ganesha.nfsd -f /etc/ganesha/ganesha.conf -F 2016-11-02 14:12:19.516992 7ffb60f880c0 -1 auth: unable to find a keyring on /var/lib/ceph/radosgw/ceph-rgw.magna104/keyring: (2) No such file or directory 2016-11-02 14:12:19.519479 7ffb60f880c0 -1 monclient(hunting): authenticate NOTE: no keyring found; disabled cephx authentication 2016-11-02 14:12:19.520188 7ffb60f880c0 -1 Couldn't init storage provider (RADOS) I think this is likely the root cause of the inability to use the cluster, because although the ganesha.conf is messy (has a lot of commented-out sections from a prior glusterfs setup), it looks clear that the RGW FSAL has initialized. I switched to using a config file based on my own template (ganesha.conf.matt), which among other things has: 1. settings to bypass the idmapper (new in 2.4.0): Allow_Numeric_Owners = true; Only_Numeric_Owners = true; 2. moves logging to /tmp/ganesha-rgw.log (not for production) Reviewing the log, I now see: 1. the RGW fsal is found and nfs-ganesha tries to intialize it 2. the RGW module init function fails (due to the aforementioned lack of a keyring, or at any rate, failure to init RADOS, which almost certainly is due to misconfiguration of ceph.conf or related files on the host 3. the subsequent report of now export entries found is actually a side-effect of problem #2--i.e., it is telling us that we failed to initialize the FSAL driver for the single configured export entry 02/11/2016 14:18:00 : epoch 5819f597 : magna105 : ganesha.nfsd-29798[main] load\ _fsal :NFS STARTUP :DEBUG :Loading FSAL RGW with /usr/lib64/ganesha/libfsalrgw.\ so 02/11/2016 14:18:00 : epoch 5819f597 : magna105 : ganesha.nfsd-29798[main] init\ :FSAL :DEBUG :RGW module registering. 02/11/2016 14:18:00 : epoch 5819f597 : magna105 : ganesha.nfsd-29798[main] init\ _config :FSAL :DEBUG :RGW module setup. 02/11/2016 14:18:00 : epoch 5819f597 : magna105 : ganesha.nfsd-29798[main] crea\ te_export :FSAL :CRIT :RGW module: librgw init failed (-5) 02/11/2016 14:18:00 : epoch 5819f597 : magna105 : ganesha.nfsd-29798[main] fsal\ _put :FSAL :INFO :FSAL RGW now unused 02/11/2016 14:18:00 : epoch 5819f597 : magna105 : ganesha.nfsd-29798[main] pseu\ dofs_create_export :FSAL :DEBUG :Created exp 0x7f7eeb7056d0 - / 02/11/2016 14:18:00 : epoch 5819f597 : magna105 : ganesha.nfsd-29798[main] main\ :NFS STARTUP :WARN :No export entries found in configuration file !!! 02/11/2016 14:18:00 : epoch 5819f597 : magna105 : ganesha.nfsd-29798[main] conf\ ig_errs_to_log :CONFIG :CRIT :Config Fi In fine: I'll try to fix the Ceph setup. But as background, the nfs-ganesha running the RGW fsal contains an embedded instance of Ceph's radosgw--and for that reason, it needs all of the supporting configuration (ceph.conf, keyrings, etc) that are needed on an radosgw node. (One easy way to test, therefore, is to set up nfs-ganesha with FSAL RGW on an existing, and already tested radosgw instance in the cluster, rather than on a separate node.) I can configure the NFS with rgw without any issues, Because of config and setup related issues above problem observed in setup. Moving this to NOTABUG and CLOSED state. |