Back to bug 1581337

Who When What Removed Added
Nir Magnezi 2018-05-22 14:29:48 UTC Target Release --- 13.0 (Queens)
Target Milestone --- zstream
Red Hat Bugzilla Rules Engine 2018-05-22 14:29:54 UTC Keywords ZStream
Target Release 13.0 (Queens) ---
Assaf Muller 2018-05-22 14:37:45 UTC Keywords Triaged
Priority unspecified high
Status NEW ASSIGNED
Assignee amuller nmagnezi
Red Hat Bugzilla Rules Engine 2018-05-22 15:13:47 UTC Target Release --- 13.0 (Queens)
Nir Magnezi 2018-06-12 09:33:15 UTC Doc Text Cause:
In order to use ping type health monitor the haproxy (default software we use in our driver for network load balancing) version must be at least 1.6

Consequence:
Any use of an older haproxy version, made the health-check to be TCP connect without the user's knowledge.

The upstream community fixed that by adding a check in the code, that determine the haproxy version that is in use and acts accordingly:
If haproxy version >= 1.6, we can use ping.
Else, we keep using TCP connect (in the absence of other solution for those haproxy versions, it is better to do so rather than breaking it altogether).

The problem we have in OSP13 GA is that we ship haproxy as a part of RHEL channels, which uses an ancient version of haproxy. Thus, when OSP13 users configure ping type health monitor, they will get TCP connect instead.

Fix:
We intend to add a config option which defaults to = block ping type health monitors in Octavia, more here[3]. We also plan to add a warning about this to raise both user and operator awareness.
Doc Type If docs needed, set a value Bug Fix
Nir Magnezi 2018-06-12 11:27:16 UTC Status ASSIGNED ON_DEV
Nir Magnezi 2018-06-19 11:36:43 UTC Doc Type Bug Fix Known Issue
Sandra McCann 2018-06-27 15:54:24 UTC CC samccann
Doc Text Cause:
In order to use ping type health monitor the haproxy (default software we use in our driver for network load balancing) version must be at least 1.6

Consequence:
Any use of an older haproxy version, made the health-check to be TCP connect without the user's knowledge.

The upstream community fixed that by adding a check in the code, that determine the haproxy version that is in use and acts accordingly:
If haproxy version >= 1.6, we can use ping.
Else, we keep using TCP connect (in the absence of other solution for those haproxy versions, it is better to do so rather than breaking it altogether).

The problem we have in OSP13 GA is that we ship haproxy as a part of RHEL channels, which uses an ancient version of haproxy. Thus, when OSP13 users configure ping type health monitor, they will get TCP connect instead.

Fix:
We intend to add a config option which defaults to = block ping type health monitors in Octavia, more here[3]. We also plan to add a warning about this to raise both user and operator awareness.
In order to use the PING type health monitor, the HAProxy (default software we use in our driver for network load balancing) version must be at least 1.6. Any use of an older HAProxy version makes the health-check be TCP connect without the user's knowledge.

The upstream community fixed that by adding a check in the code, that determine the HAProxy version that is in use and acts accordingly:
If HAProxy version 1.6 or later, we can use PING.
Otherwise, we keep using TCP connect (in the absence of any other solution for those haproxy versions, it is better to do so rather than breaking it altogether).

The problem we have in OSP13 GA is that we ship HAProxy as a part of RHEL channels, which uses an old version of HAProxy. Thus, when OSP13 users configure the PING type health monitor, they will get TCP connect instead.
Sandra McCann 2018-06-28 16:04:02 UTC CC samccann
Takuya Ito 2018-07-03 12:14:32 UTC CC takito
Carlos Goncalves 2018-07-11 10:00:08 UTC CC cgoncalves
Scott Lewis 2018-07-26 16:24:06 UTC Target Milestone zstream ---
Carlos Goncalves 2018-10-31 19:39:38 UTC Blocks 1644883
Carlos Goncalves 2018-10-31 19:41:23 UTC Blocks 1644883
Depends On 1644883
Carlos Goncalves 2019-01-23 10:43:04 UTC Status ON_DEV POST
Target Milestone --- z5
PnT Account Manager 2019-02-14 15:02:00 UTC CC nyechiel
Carlos Goncalves 2019-02-18 13:03:42 UTC Target Milestone z5 z6
Nir Magnezi 2019-04-08 11:46:43 UTC Status POST MODIFIED
Fixed In Version openstack-octavia-2.0.4-3.el7ost
errata-xmlrpc 2019-04-08 14:30:38 UTC Status MODIFIED ON_QA
Steve Linabery 2019-04-09 20:02:31 UTC Status ON_QA MODIFIED
Jon Schlueter 2019-04-11 12:04:54 UTC Status MODIFIED ON_QA
Nir Magnezi 2019-04-16 11:39:11 UTC Blocks 1700359
Bruna Bonguardo 2019-04-23 11:37:35 UTC CC bbonguar
Depends On 1698860
QA Contact astafeye bbonguar
Steve Linabery 2019-04-29 17:12:30 UTC Keywords TestOnly
CC slinaber
Bruna Bonguardo 2019-05-01 08:34:38 UTC Status ON_QA ASSIGNED
Nir Magnezi 2019-05-01 12:41:09 UTC Flags needinfo?(bbonguar)
Nir Magnezi 2019-05-02 08:21:20 UTC Status ASSIGNED ON_QA
Bruna Bonguardo 2019-05-02 10:41:16 UTC Status ON_QA VERIFIED
Flags needinfo?(bbonguar)
Shelley Dunne 2019-06-21 19:50:12 UTC CC shdunne
Target Milestone z6 z7
mgeary 2019-07-01 09:56:41 UTC CC mgeary
Doc Text In order to use the PING type health monitor, the HAProxy (default software we use in our driver for network load balancing) version must be at least 1.6. Any use of an older HAProxy version makes the health-check be TCP connect without the user's knowledge.

The upstream community fixed that by adding a check in the code, that determine the HAProxy version that is in use and acts accordingly:
If HAProxy version 1.6 or later, we can use PING.
Otherwise, we keep using TCP connect (in the absence of any other solution for those haproxy versions, it is better to do so rather than breaking it altogether).

The problem we have in OSP13 GA is that we ship HAProxy as a part of RHEL channels, which uses an old version of HAProxy. Thus, when OSP13 users configure the PING type health monitor, they will get TCP connect instead.
HAProxy, used for network load balancing, must be version 1.6 or higher to correctly support the PING type health monitor.

However the version of HAProxy included Red Hat OpenStack Platform 13 is an older version than 1.6, that uses TCP connect instead when you configure the PING type health monitor.
mgeary 2019-07-08 15:53:21 UTC Doc Text HAProxy, used for network load balancing, must be version 1.6 or higher to correctly support the PING type health monitor.

However the version of HAProxy included Red Hat OpenStack Platform 13 is an older version than 1.6, that uses TCP connect instead when you configure the PING type health monitor.
HAProxy, used for network load balancing, must be version 1.6 or higher to correctly support the PING type health monitor.

The version of HAProxy included with Red Hat OpenStack Platform 13 is an older version than 1.6 that uses TCP connect instead when you configure the PING type health monitor.
errata-xmlrpc 2019-07-09 18:33:02 UTC Status VERIFIED RELEASE_PENDING
errata-xmlrpc 2019-07-10 13:01:59 UTC Status RELEASE_PENDING CLOSED
Resolution --- ERRATA
Last Closed 2019-07-10 13:01:59 UTC
errata-xmlrpc 2019-07-10 13:02:17 UTC Link ID Red Hat Product Errata RHBA-2019:1744
Nir Magnezi 2019-09-10 14:09:36 UTC CC nmagnezi
Carlos Goncalves 2020-02-04 13:59:40 UTC Blocks 1798064

Back to bug 1581337