Alon Bar-Lev has posted comments on this change.

Change subject: bootstrap: new implementation for apache-sshd usage
......................................................................


Patch Set 1: Verified

Hello all,

This is a complete re-write of SSH handling during various of bootstraps.

It was tested using the following sequences:

1. Standard engine, bootstrap from engine.

2. ovirt-node (upstream), registration from node. Needs some tweaks, like 
restarting the vdsm-reg but as far as the sequence is concerned it is working.

3. ovirt-node (upstream), bootstrap from engine, works although after bootstrap 
it returns to "accept" requirement as if node was registered, if this is not 
the correct behaviour I need to know. Nothing fails so I guess this is how it 
is working.

4. ovirt-node (downstream, rhevh-6.3-20120710.0.el6_3.iso), registration from 
node, working if creating 3.0 domain, and  after boot or registration manually 
renaming bridge name to ovirtmgmt.

5. ovirt-node (downstream, rhevh-6.3-20120710.0.el6_3.iso) behaves the same as 
upstream.

6. ovirt-node (downstream, rhevh-6.3-20120710.0.el6_3.iso) reinstall works see 
comments.

---

As there is no actual working setup for ovirt-node, it is very difficult to 
know for 100% if the rewrite is actually OK. However, the state of the code 
level is much better and the unit tests are more complete, so fixing any issue 
which may raise is simpler.

One note though... the rhevh-6.3-20120710.0.el6_3.iso image reboots the machine 
during upgrade using:

 ---
    echo "* * * * * /sbin/reboot" > /var/spool/cron/root
    # ensure cron get's started
    crond

    echo "Finalizing Install and Rebooting (this may take a minute)"
    sleep 60
 ---

As the script never ends, reboot is done while ssh session is established. In 
this case the client side is not aware of session closure and waits until time 
out.

As we cannot modify the ovirt-functions we have can reduce the time out of the 
session.

Future nodes use plain reboot[1], which should make it work better if nothing 
still waits for the result. I did not find in history of upstream when this was 
changed.

Best practice should be to return success to server and reboot in the 
background.

Alon.

[1] 
http://gerrit.ovirt.org/gitweb?p=ovirt-node.git;a=blob;f=scripts/ovirt-functions;h=554aa80c9a0dff90a0ef298948c084e136b87591;hb=HEAD#l905

--
To view, visit http://gerrit.ovirt.org/6722
To unsubscribe, visit http://gerrit.ovirt.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I50ba60f2db364114907485da3074feb714615e0c
Gerrit-PatchSet: 1
Gerrit-Project: ovirt-engine
Gerrit-Branch: master
Gerrit-Owner: Alon Bar-Lev <alo...@redhat.com>
Gerrit-Reviewer: Alon Bar-Lev <alo...@redhat.com>
Gerrit-Reviewer: Barak Azulay <bazu...@redhat.com>
Gerrit-Reviewer: Doron Fediuck <dfedi...@redhat.com>
Gerrit-Reviewer: Juan Hernandez <juan.hernan...@redhat.com>
Gerrit-Reviewer: Ofer Schreiber <oschr...@redhat.com>
Gerrit-Reviewer: Yair Zaslavsky <yzasl...@redhat.com>
_______________________________________________
Engine-patches mailing list
Engine-patches@ovirt.org
http://lists.ovirt.org/mailman/listinfo/engine-patches

Reply via email to