Too many ephemeral ports open on postgres container
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Invalid
|
Medium
|
Fuel Library (Deprecated) | ||
6.0.x |
Won't Fix
|
Medium
|
Matthew Mosesohn | ||
7.0.x |
Invalid
|
Medium
|
Charlie Ott | ||
8.0.x |
Invalid
|
Medium
|
Fuel Library (Deprecated) |
Bug Description
While trying to rebuild my postgres container and restore from pgsql/9.3 data directory from a previously lost container(http://
The TCP connection never happens and is timed out over and over. Thus I cannot log in to the fuel UI:
2015-04-06 17:29:29.881 868 TRACE keystone.
2015-04-06 17:29:29.881 868 TRACE keystone.
2015-04-06 17:29:29.881 868 TRACE keystone.
So I checked the container status with docker:
[root@fuel ~]# docker ps -a | grep postgres | grep Up
d78296d733f2 fuel/postgres_
and it looks good, so then looking at netstat inside the container:
dockerctl shell postgres su postgres -c "netstat -a"
Shows that there are 748 waiting connections on my instance. (Currently)
This seems very excessive and inefficient.
using openstack paste, as this output is very long and will make this ticket ridiculous to scroll through: http://
ps -ef inside the container shows postmaster running and listening on the right port... but running it multiple times shows another instance starting/stopping over and over...???
[root@fuel ~]# dockerctl shell postgres su postgres -c "ps -ef | grep pgsql"
root 2895 1 0 19:14 ? 00:00:00 sudo -u postgres /usr/pgsql-
postgres 2896 2895 9 19:14 ? 00:02:20 /usr/pgsql-
root 11654 0 0 19:37 ? 00:00:00 su postgres -c ps -ef | grep pgsql
postgres 11661 11654 0 19:37 ? 00:00:00 bash -c ps -ef | grep pgsql
postgres 11663 11661 0 19:37 ? 00:00:00 grep pgsql
postgres 11664 2896 0 19:37 ? 00:00:00 /usr/pgsql-
[root@fuel ~]# dockerctl shell postgres su postgres -c "ps -ef | grep pgsql"
root 2895 1 0 19:14 ? 00:00:00 sudo -u postgres /usr/pgsql-
postgres 2896 2895 9 19:14 ? 00:02:30 /usr/pgsql-
root 4732 0 0 19:39 ? 00:00:00 su postgres -c ps -ef | grep pgsql
postgres 4738 4732 0 19:39 ? 00:00:00 bash -c ps -ef | grep pgsql
postgres 4741 4738 0 19:39 ? 00:00:00 grep pgsql
[root@fuel ~]# dockerctl shell postgres su postgres -c "ps -ef | grep pgsql"
root 2895 1 0 19:14 ? 00:00:00 sudo -u postgres /usr/pgsql-
postgres 2896 2895 9 19:14 ? 00:02:31 /usr/pgsql-
root 6480 0 0 19:39 ? 00:00:00 su postgres -c ps -ef | grep pgsql
postgres 6487 6480 0 19:39 ? 00:00:00 bash -c ps -ef | grep pgsql
postgres 6489 6487 0 19:39 ? 00:00:00 grep pgsql
postgres 6490 2896 0 19:39 ? 00:00:00 /usr/pgsql-
description: | updated |
description: | updated |
Changed in fuel: | |
status: | New → Confirmed |
importance: | Undecided → High |
assignee: | nobody → Matthew Mosesohn (raytrac3r) |
tags: | added: customer-found |
tags: | added: qa-agree-7.0 release-notes |
no longer affects: | fuel/6.1.x |
Changed in fuel: | |
status: | Confirmed → Won't Fix |
Changed in fuel: | |
status: | Incomplete → Won't Fix |
tags: | added: area-library |
Changed in fuel: | |
assignee: | Charlie Ott (charlieott) → Fuel Library Team (fuel-library) |
milestone: | 6.1 → 8.0 |
this is on a fuel 6.0 instance i have been upgrading since 5.0, (approx 8 months)