'ntpq: read: Connection refused' on controllers while performing any query to NTPD from vrouter namespace using 'ntpq'
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Fix Released
|
High
|
Dennis Dmitriev |
Bug Description
Reproduced on CI (for example: http://
, in the following system tests:
ha_
check_
ha_
backup_
manual_
auto_
negative_
negative_
external_ntp_ha
external_dns_ha
On Ubuntu controllers in any cluster configuration, the following error could appear:
root@node-1:~# ip netns exec vrouter ntpq -p
ntpq: read: Connection refused
This is happened because 'ntpq' is connected to NTPD using IPv6, but started by OCF script NTPD is using '-4' option and rejects all IPv6 requests.
Possible solutions:
- remove '-4' option from OCF script ns_ntp. This option is not used for other nodes in cluster;
- fix system tests and reflect in documentation that user use 'ntpq' by manually providing IPv4 address:
remote refid st t when poll reach delay offset jitter
ISO version: {u'build_id': u'2015-
description: | updated |
Changed in fuel: | |
assignee: | Fuel Library Team (fuel-library) → Dennis Dmitriev (ddmitriev) |
status: | New → In Progress |
Changed in fuel: | |
status: | In Progress → Fix Committed |
Changed in fuel: | |
status: | Fix Committed → Fix Released |
I actually wonder why we use some ipv6 stuff if we don't have any settings and mentions about it in Fuel. So, listening on ipv4 only seems good to me, cause there is no support of ipv6 in whole product.
In other side, enabling ipv6 is not a big problem - we just should do things carefully and look for possible problems with it.