Fix error check for rsync operation in fuel-createmirror

Bug #1510976 reported by Sergey Kulanov on 2015-10-28
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Fuel for OpenStack
High
Sergey Kulanov
7.0.x
High
Sergey Kulanov

Bug Description

We need to fix error check for rsync operation

https://github.com/openstack/fuel-mirror/blob/master/deb-mirror#L523-L529

since for now we are checking exit code for find+chmod operation instead of rsync

https://github.com/openstack/fuel-mirror/blob/master/deb-mirror#L523-L529

Changed in fuel:
assignee: nobody → Vitaly Parakhin (vparakhin)
status: New → Confirmed
Changed in fuel:
assignee: Vitaly Parakhin (vparakhin) → Sergey Kulanov (skulanov)
status: Confirmed → In Progress

Reviewed: https://review.openstack.org/239909
Committed: https://git.openstack.org/cgit/openstack/fuel-mirror/commit/?id=5a902c79beb5cb4e3c53b9d13668d6bed7e0552d
Submitter: Jenkins
Branch: master

commit 5a902c79beb5cb4e3c53b9d13668d6bed7e0552d
Author: Vitaly Parakhin <email address hidden>
Date: Wed Oct 28 13:16:49 2015 +0200

    Fix the setting of pool files permissions

    Fix was incorrectly placed between main rsync and status code
    check, so rearranging it properly.

    Closes-bug: #1510976
    Change-Id: I37892d7803a8590d643ea8fc37f525fd7637a997

Changed in fuel:
status: In Progress → Fix Committed
Roman Vyalov (r0mikiam) on 2015-11-05
tags: added: area-build
Sergey Kulanov (skulanov) wrote :

Team, we have fix for a long time, please check

Denis Meltsaykin (dmeltsaykin) wrote :

What is the impact? It's not clear what is broken and why it needs to be fixed.

Sergey Kulanov (skulanov) wrote :

@Denis, the issue is that we incorrectly detect exit code, since we always check `find` command result instead of `rsync` command result.

The impact is that if we have non-zero exit code for `rsync` command we'll never catch this because we have `find`.

In short incorrect flow:
1) rsync
2) find
3) check_rsync_code

but MUST be:

1) rsync
2) check_rsync_code
3) find

Denis Meltsaykin (dmeltsaykin) wrote :

Sergey, the question was about impact, i.e. how the issue affects a workflow? I'm just not sure that it is of High importance. It looks like there was no real life failure which led to broken workloads, right?

Change abandoned by Sergey Kulanov (<email address hidden>) on branch: stable/7.0
Review: https://review.openstack.org/240175

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers