Hide Forgot
Description of problem: Further to VM SSA, evm.log contain this error: [----] E, [2016-12-06T08:54:39.865473 #18632:f3514c] ERROR -- : MIQ(ManageIQ::Providers::Redhat::InfraManager::Host#connect_ssh) SSH connection failed for [seal10.qa.la b.tlv.redhat.com] with [ArgumentError: invalid option(s): attempt] masayag: It seems that the SSH errors in the code, Are regression due to the following commit [1]: [1] https://github.com/ManageIQ/manageiq/pull/11964 commit 2e8f5985e61d1566cca42649cd7c0b0f21fccdec Author: Joe Rafaniello <jrafanie> Date: Fri Oct 14 18:50:43 2016 -0400 Retry retryable authentication failures Version-Release number of selected component (if applicable): CFME-5.7.0.13 (Seen on RHV-4.0.5) Additional info: Not sure as for the impact this SSH problem may cause.
Can you provide a full backtrace or an example of how to recreate this error? What ovirt/rhev provider code are you seeing this with? I searched for callers to authentication_check* methods to make sure this wasn't a breaking change but maybe I missed one. Thanks!
Created attachment 1228616 [details] evm.log In this evm.log, For VM SSA, done at 2016-12-06T10:43:52 can see this error in evm.log, few minutes after: ERROR -- : MIQ(ManageIQ::Providers::Redhat::InfraManager::Host#connect_ssh) SSH connection failed for [seal10.qa.lab.tlv.redhat.com] with [ArgumentError: invalid option(s): attempt]
Joe, In order to recreate this error, need to run VM SSA. I did it on RHV, but code indicate it's a core bug, so it may occur for other providers as well, like VMWare. evm.log has been attached. As for your question in the code, masayag is the one who tracked it: Moti, Would you please address Joe's question in comment 2? Thanks.
Thank you for finding this Ilanit. Can you try this PR? I was able to recreate the problem and put a pretty ugly but working unit test on it. https://github.com/ManageIQ/manageiq/pull/13026 Note, you can use this to get a patch file: https://github.com/ManageIQ/manageiq/pull/13026.patch
Tested CFME-5.7.0.13 + https://github.com/ManageIQ/manageiq/pull/13026.patch (Thanks masayag, for helping with adding the patch): Same CFME-5.7.0.13, where bug was reported, Running VM SSA for the same VM, that was run before. Now there is no ERROR in the evm.log, 15 min after the SSA operation. Therefore, this patch is verified. Thank you very much for your quick response.
Can we push that to 5.7.0?
New commit detected on ManageIQ/manageiq/master: https://github.com/ManageIQ/manageiq/commit/ab40ebedc07c7ebd06ac4d9f759cafada5bbfdfb commit ab40ebedc07c7ebd06ac4d9f759cafada5bbfdfb Author: Joe Rafaniello <jrafanie> AuthorDate: Tue Dec 6 12:51:43 2016 -0500 Commit: Joe Rafaniello <jrafanie> CommitDate: Tue Dec 6 13:00:38 2016 -0500 Filter attempt from the authentication_check options Various implementations of verify_credentials aren't expecting an :attempt => 1 hash key/value. It's only used at a higher layer so exclude it from the authentication_check call. https://bugzilla.redhat.com/show_bug.cgi?id=1402035 This was a regression introduced in #11964 app/models/mixins/authentication_mixin.rb | 2 +- spec/models/mixins/authentication_mixin_spec.rb | 6 ++++++ 2 files changed, 7 insertions(+), 1 deletion(-)
> Can we push that to 5.7.0? Oved, this has been marked as a blocker for backport to euwe and 5.7. Your team found this JUST in time... a few days later and it might not have made it. Thanks for finding this.
Thanks for the fast response and for pushing this to 5.7.0. Ilanit, thanks indeed for finding this.
*** Bug 1392460 has been marked as a duplicate of this bug. ***
Verified on CFME-5.8.0.13/RHV-4.1.1.8 VM SSA for NFS disk worked OK. There was no error in evm.log