Red Hat Bugzilla – Bug 983051
[RHSC] Import cluster when one of the peers is unresolvable or unreachable from the engine, fails.
Last modified: 2015-05-13 12:27:59 EDT
Created attachment 771561 [details]
Description of problem:
Server A has 2 peers, B and C.
B was peer probed from A using the IP address, C was peer probed from A, using the hostname known only among the three servers, by means of entries in /etc/hosts.
When trying to import the cluster using the Console, using the IP address of A, the import fails with the following exception -
Another case when the import fails, is when one of the hosts is down, in other words, unreachable. In this case, the import fails with the following exception-
org_apache_sshd_common_RuntimeSshException: Failed to get the session_
Version-Release number of selected component (if applicable):
Red Hat Storage Console Version: 2.1.0-0.bb5.el6rhs
Steps to Reproduce:
The following are the steps to reproduce the first case described above -
1. Add entries for servers A and B in /etc/hosts of C, to identify them as host1 and host2 respectively and similar entries in A and B too.
2. Run the following on A -
# gluster peer probe <IP-address-of-B>
# gluster peer probe host3
host3 is to identify server C.
3. Now, try to import this cluster via the Console, by providing the IP address of A, in the import cluster screen.
For the second case described above,
1. Run 'gluster peer probe' from server A to B and C using the IP addresses of B and C.
2. Import the cluster via the Console, using the IP address of A.
Import fails with the above described exceptions.
For both the above described cases, import should succeed for those hosts that are reachable and resolvable. For others alone, it should fail.
Attaching engine logs and vdsm logs from server A.
Created attachment 771563 [details]
For Case A:
You have to add entries in engine host file also
Then only the engine can reach the host using the IP address
given in the host file.
For Case B:
Currently the engine gets the peer status from (any of the) given
server in the Gluster and will send a call to add them individually;
and if any peer (server) is unreachable by the engine, it fails.
Patch sent to ovirt-engine upstream:
Verified as fixed in Red Hat Storage Console Version: 2.1.2-0.21.beta1.el6_4. Import cluster works in both cases mentioned above.
Please review the edited DocText and signoff.
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.