Commit Graph

1431 Commits

Author SHA1 Message Date
Oleksii Kliukin 2e226dee26 Avoid overwriting infrastrure roles.
When a role is defined in the infrastructure roles and the cluster
manifest use the infrastructure role definition and add flags
defined in the manifest.

Previously the role has been overwritten by the definition from the
manifest.  Because a random password is generated for each role from the
manifest the applications relying on the infrastructure role credentials
from the infrastructure roles secret were unable to connect.
2017-12-04 10:49:25 +01:00
Oleksii Kliukin dd0affc390 Tweak our reaction to the cluster upgrade process.
Previously, the operator started to move the pods off the nodes to be
decomissioned by watching the eol_node_label value. Every new postgres
pod has been created with the anti-affinity to that label, making sure
that the pods being moved won't land on another to be decomissioned
node.

The changes introduce another label that indicates the ready node.  The
new pod affinity will esnure that the pod is only scheduled to the node
marked as ready, discarding the previous anti-affinity.  That way the
nodes can transition from the pending-decomission to the other statuses
(drained, terminating) without having pods suddently scaled to them.

In addition, rename the label that triggers the start of the upgrade
process to node_eol_label (for consistency with node_readiness_label)
and set its default vvalue to lifecycle-status:pending-decomission.
2017-11-30 14:11:49 +01:00
Manuel Gómez a39e89d155
Merge pull request #170 from zalando-incubator/bugfix/connection_leak_and_sync_role_options
Fix the connection leak and user options sync.
2017-11-28 11:10:16 +01:00
Oleksii Kliukin 1ffe98ba9f Fix the connection leak and user options sync.
- fix the lack of closing the cursor for the query that returned no
rows.
- fix syncing of the user options, as previously those were not
  fetched from the database.
2017-11-27 16:46:34 +01:00
Oleksii Kliukin 68bb3cd52d Add an example of the new parameter with multiple values. 2017-11-22 10:43:35 +01:00
Oleksii Kliukin 086ead03f5 Warn about attempts to use escape quotes. 2017-11-22 10:43:35 +01:00
Oleksii Kliukin f6a2225c38 rename the parameter in the README. 2017-11-22 10:43:35 +01:00
Oleksii Kliukin 975b21f633 Rename api roles configuration parameter.
Change api_roles_configuration to team_api_role_configuration
2017-11-22 10:43:35 +01:00
Oleksii Kliukin 6b2f5071f7 Special case for search_path in user options.
- search_path accepts a list of values that cannot be quoted, as
  quoting would make PostgreSQL interpret the result as a single
  value. Since we require quoting of values with commas in the
  operator's configMap in order to avoid confusing them with the
  separate map entities, we need to strip those quotes before
  passing the value to PostgreSQL.
- make ftm run
2017-11-22 10:43:35 +01:00
Oleksii Kliukin 2079d811b4 Add tests for the string splitting function. 2017-11-22 10:43:35 +01:00
Oleksii Kliukin e95f80e351 Make configMap marshaling code aware of quotes.
A value in a configMap that is a map itself
(a key:value string separated by commas) may include commans inside
quotes (i.e. search_path:"public,"$user"). The changes make marshaling
code process such cases correctly.
2017-11-22 10:43:35 +01:00
Oleksii Kliukin 2352fc9a39 go fmt run 2017-11-22 10:43:35 +01:00
Oleksii Kliukin 71f57c9fe3 Fix escaping of parameter values and extra spaces.
- document the newly introduced option (for now in the main README)
- make query error output more readable.
2017-11-22 10:43:35 +01:00
Oleksii Kliukin 415a7fdc4d Allow global configuration options for API roles.
Add options to the PgUser structure, potentially allowing to set
per-role options in the cluster definition as well.

Introduce api_roles_configuration operator option with the default
of log_statement=all
2017-11-22 10:43:35 +01:00
Oleksii Kliukin 6dcd074ea0 Allow per-cluster setting of a docker image.
Add dockerImage cluster configuration parameter that overrides global
operator defaults when set to a non-empty value.
2017-11-14 11:53:04 +01:00
Oleksii Kliukin c25e849fe4 Fix a failure to create new statefulset at sync.
Also do a fmt run.
2017-11-08 18:24:17 +01:00
Andras Vaczi 6813bf5921
Merge pull request #160 from zalando-incubator/readme-stackenblocken
Stackeblockenize the first section of the readme
2017-11-08 11:38:55 +01:00
András Váczi 8c185e407c Stackeblockenize the first section of the readme 2017-11-08 11:04:56 +01:00
Andras Vaczi 52d7cda186
Merge pull request #159 from zalando-incubator/link-patroni
link to Patroni
2017-11-08 10:07:56 +01:00
Henning Jacobs 3cf68afeb5
link to Patroni 2017-11-07 22:54:50 +01:00
Murat Kabilov 86803406db
use sync methods while updating the cluster 2017-11-03 12:00:43 +01:00
Georg Kunz 47dd766fa7 Add node toleration config to PodSpec (#151)
* Add node toleration config to PodSpec

This allows to taint nodes dedicated to Postgres and prevents other pods from running on these nodes.

* Document taint and toleration setup

And remove setting from default operator ConfigMap

* Allow to overwrite tolerations with Postgres manifest
2017-11-02 19:10:44 +01:00
Oleksii Kliukin ce960e892a
Create new databases and change owners of existing ones during sync. (#153)
* Create new databases and change owners of existing ones during sync.
2017-11-02 17:46:33 +01:00
zerg-junior d3679bfd4a Merge pull request #150 from zerg-junior/bugfix/readme
Change cluster name to "acid-minimal" in readme
2017-10-25 11:15:11 +02:00
zerg-junior d06415c6a5 Change cluster name to "acid-minimal" in readme
Cluster created via `minimal-postgres-manifest` has the name `acid-minimal-cluster`, not `acid-test-cluster`
2017-10-25 10:40:31 +02:00
Jan Mussler a98a7c95c2 Reorganize Readme (#142)
removing parts of config.

* chaning secret name pattern to make things shorter.

* Move section on self building docker image.

* Fix typo.

* Bump image.

* bump version for pdb fix.

* Changes in regards to review.

* Fix xhyve driver link.

* Move to new api, remove service account, not needed for minikube.

* Changed minimal manifest and example to use right file.

* Added service account for operator again, it is needed in pods anyways later.
2017-10-24 20:42:22 +02:00
Oleksii Kliukin 7a76be7d3e Minor fixes around PDB (pod-distruption-budget) syncing: (#147)
- Call comparison function in the case of the sync as well as for update
- Include full cluster name in PDB name
- Assign cluster labels to the PDB object
2017-10-23 12:26:59 +02:00
Murat Kabilov c17aabb642 fix pod disruption budget labels (#146) 2017-10-20 15:01:51 +02:00
Alexander Kukushkin 39200ba8d4 Enable k8s leader election (#145)
and bump docker image version
2017-10-20 13:58:15 +02:00
Murat Kabilov 661b141849 Fix Pod Disruption Budget null pointer exception 2017-10-20 11:43:50 +02:00
Murat Kabilov a1deae198b add missing master matchLabel for the PDB (#144) 2017-10-20 11:26:40 +02:00
Alexander Kukushkin a98c712a52 Change spilo docker image to demospilo (#141)
Image size is slightly more than 24MB, it doesn't contain wal-e and not suitable for production, but it is very good for demo purposes.
2017-10-19 13:53:12 +02:00
Oleksii Kliukin eba23279c8 Kube cluster upgrade 2017-10-19 10:49:42 +02:00
Oleksii Kliukin 1dbf259c76 Retry opening DB connections. (#140)
Make sure DB connection retry also reopens a connection after closing it
2017-10-18 16:28:00 +02:00
Oleksii Kliukin 99870d8eac Fix division by zero when connecting to the DB.
Apparently the retry function's first parameter is the duration of
a single attempt and it cannot be zero.
2017-10-18 10:44:49 +02:00
Murat Kabilov 202f2de988 Retry connecting to pg 2017-10-17 17:03:50 +02:00
Murat Kabilov 6c4cb4e9da Perform manual failover during the scale down 2017-10-16 17:41:23 +02:00
Murat Kabilov 5b29576a8e Remove redundant constants 2017-10-16 15:52:48 +02:00
Jan Mussler 959b483d11 Update README.md (#133) 2017-10-12 18:24:10 +02:00
Murat Kabilov 3b32265258 Set status of the cluster on sync fail/success 2017-10-12 15:10:42 +02:00
Jan Mussler cec695d48e Superuser toggle for team members
Make superuser toggleable for team members. Add and "admin" role to team members if superuser is disabled.
2017-10-12 15:01:54 +02:00
Murat Kabilov 702d901bd9 use clear name for env var denoting namespace to watch (#129) 2017-10-12 10:42:20 +02:00
Murat Kabilov 8d5faaa5a5 return idle status when worker has nothing to do 2017-10-11 15:42:20 +02:00
Oleksii Kliukin 793defef72 Fix pod wait timeouts.
Previously, a timer had been reset on every message received through
the pod channel.
2017-10-11 14:58:37 +02:00
Murat Kabilov 83c8d6c419 Extend diagnostic api with worker status info 2017-10-11 12:26:09 +02:00
Murat Kabilov 4bc2284b57 bump go version to 1.9 2017-10-11 12:01:35 +02:00
Oleksii Kliukin 846d919235 Fix the endpoints (#121) 2017-10-10 10:03:07 +02:00
Oleksii Kliukin b8b13c569b Describe available endpoints. (#120) 2017-10-09 18:07:18 +02:00
Murat Kabilov 32aa7270e6 Use round-robin strategy while assigning workers 2017-10-09 16:56:27 +02:00
Jan Mussler d6393d46cb Merge pull request #117 from zalando-incubator/crd
Replace TPR with CRD
2017-10-09 12:23:07 +02:00