How to Clean Up Deleted Projects in Cloudera Data Science Workbench

We have got a few customer complains about the fact that currently Cloudera Data Science Workbench (CDSW) does not release the underlining project files on disk after the project is deleted from within the CDSW web console.

This can be re-produced easily by creating a dummy project in CDSW, check the project directory created under /var/lib/cdsw/current/projects/projects/0, and then delete the project again, you will see that the newly created project directory is not removed on the file system.

This is a known bug and reported internally in Cloudera, however, it has not been fixed as yet.

To workaround the issue, you can setup a simple shell script to detect the orphaned project and remove it automatically.

Steps as below:

1. Get the list of Project IDs from the directory /var/lib/cdsw/current/projects/projects/0/ on the master host, in my example, it returned below:

ls /var/lib/cdsw/current/projects/projects/0/
1  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48

2. Run ‘cdsw status’ command on the master host to capture the DB pod id:

Sending detailed logs to [/tmp/cdsw_status_T8jRig.log] ...
CDSW Version: [1.3.0:9bb84f6]
OK: Application running as root check
OK: Sysctl params check
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|                 NAME                |   STATUS   |           CREATED-AT          |   VERSION   |   EXTERNAL-IP   |          OS-IMAGE         |         KERNEL-VERSION         |   GPU   |   STATEFUL   |
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|   ericlin-xxx.xxx-1.com   |    True    |   2018-07-11 03:30:45+00:00   |   v1.6.11   |       None      |   CentOS Linux 7 (Core)   |   3.10.0-514.26.2.el7.x86_64   |    0    |    False     |
|   ericlin-xxx.xxx-2.com   |    True    |   2018-07-11 03:30:32+00:00   |   v1.6.11   |       None      |   CentOS Linux 7 (Core)   |   3.10.0-514.26.2.el7.x86_64   |    0    |     True     |
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
2/2 nodes are ready.
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|                             NAME                            |   READY   |    STATUS   |   RESTARTS   |           CREATED-AT          |       POD-IP      |      HOST-IP      |   ROLE   |
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|             etcd-ericlin-xxx.xxx-1.com                      |    1/1    |   Running   |      5       |   2018-07-11 03:31:44+00:00   |   172.26.12.157   |   172.26.12.157   |   None   |
|        kube-apiserver-ericlin-xxx.xxx-1.com                 |    1/1    |   Running   |      5       |   2018-07-11 03:30:31+00:00   |   172.26.12.157   |   172.26.12.157   |   None   |
|   kube-controller-manager-ericlin-xxx.xxx-1.com             |    1/1    |   Running   |      5       |   2018-07-11 03:31:54+00:00   |   172.26.12.157   |   172.26.12.157   |   None   |
|                  kube-dns-3911048160-30l05                  |    3/3    |   Running   |      15      |   2018-07-11 03:30:45+00:00   |    100.66.128.1   |   172.26.12.157   |   None   |
|                       kube-proxy-c4xk7                      |    1/1    |   Running   |      4       |   2018-07-11 03:30:45+00:00   |    172.26.14.58   |    172.26.14.58   |   None   |
|                       kube-proxy-k95s2                      |    1/1    |   Running   |      5       |   2018-07-11 03:30:45+00:00   |   172.26.12.157   |   172.26.12.157   |   None   |
|        kube-scheduler-ericlin-xxx.xxx-1          .com       |    1/1    |   Running   |      5       |   2018-07-11 03:31:57+00:00   |   172.26.12.157   |   172.26.12.157   |   None   |
|               node-problem-detector-v0.1-0624z              |    1/1    |   Running   |      5       |   2018-07-11 03:32:15+00:00   |   172.26.12.157   |   172.26.12.157   |   None   |
|               node-problem-detector-v0.1-b80tt              |    1/1    |   Running   |      4       |   2018-07-11 03:32:15+00:00   |    172.26.14.58   |    172.26.14.58   |   None   |
|                       weave-net-469fb                       |    2/2    |   Running   |      12      |   2018-07-11 03:30:45+00:00   |   172.26.12.157   |   172.26.12.157   |   None   |
|                       weave-net-8dzx6                       |    2/2    |   Running   |      10      |   2018-07-11 03:30:45+00:00   |    172.26.14.58   |    172.26.14.58   |   None   |
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
All required pods are ready in cluster kube-system.
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|                  NAME                  |   READY   |    STATUS   |   RESTARTS   |           CREATED-AT          |       POD-IP      |      HOST-IP      |           ROLE           |
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|         cron-1906902965-wzkp5          |    1/1    |   Running   |      2       |   2018-08-10 08:13:55+00:00   |    100.66.128.9   |   172.26.12.157   |           cron           |
|          db-1165222207-dg98q           |    1/1    |   Running   |      5       |   2018-07-11 03:32:15+00:00   |    100.66.128.6   |   172.26.12.157   |            db            |
|           engine-deps-1rvcl            |    1/1    |   Running   |      5       |   2018-07-11 03:32:15+00:00   |    100.66.128.4   |   172.26.12.157   |       engine-deps        |
|           engine-deps-njwlc            |    1/1    |   Running   |      4       |   2018-07-11 03:32:15+00:00   |     100.66.0.5    |    172.26.14.58   |       engine-deps        |
|   ingress-controller-684706958-6fzh3   |    1/1    |   Running   |      5       |   2018-07-11 03:32:14+00:00   |   172.26.12.157   |   172.26.12.157   |    ingress-controller    |
|        livelog-2502658797-kmq4l        |    1/1    |   Running   |      5       |   2018-07-11 03:32:15+00:00   |    100.66.128.3   |   172.26.12.157   |         livelog          |
|      reconciler-2738760185-1nnsp       |    1/1    |   Running   |      2       |   2018-08-10 08:13:55+00:00   |    100.66.128.2   |   172.26.12.157   |        reconciler        |
|       spark-port-forwarder-krtw6       |    1/1    |   Running   |      5       |   2018-07-11 03:32:15+00:00   |   172.26.12.157   |   172.26.12.157   |   spark-port-forwarder   |
|       spark-port-forwarder-rbhc6       |    1/1    |   Running   |      4       |   2018-07-11 03:32:15+00:00   |    172.26.14.58   |    172.26.14.58   |   spark-port-forwarder   |
|          web-3320989329-7php0          |    1/1    |   Running   |      2       |   2018-08-10 08:13:55+00:00   |    100.66.128.7   |   172.26.12.157   |           web            |
|          web-3320989329-ms63k          |    1/1    |   Running   |      5       |   2018-07-11 03:32:15+00:00   |    100.66.128.5   |   172.26.12.157   |           web            |
|          web-3320989329-zdpcj          |    1/1    |   Running   |      2       |   2018-08-10 08:13:55+00:00   |    100.66.128.8   |   172.26.12.157   |           web            |
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
All required pods are ready in cluster default.
All required Application services are configured.
All required config maps are ready.
All required secrets are available.
Persistent volumes are ready.
Persistent volume claims are ready.
Ingresses are ready.
Checking web at url: http://ericlin-xxx.xxx-1.com
OK: HTTP port check
Cloudera Data Science Workbench is ready!

You can see that from above example, the DB pod ID is: db-1165222207-dg98q

3. Run below command to connect to CDSW database:

kubectl exec db-1165222207-dg98q -ti -- psql -U sense

4. On prompt, run PostgreSQL to check which project has been delete (the one not in the DB)

sense=# SELECT id, user_id, name, slug FROM projects;
 id | user_id |      name      |      slug
----+---------+----------------+----------------
 41 |       1 | Impala Project | impala-project
 34 |       3 | Test           | test
  1 |       1 | Test           | test
 47 |      10 | tensortest     | tensortest
 44 |       9 | TestEnvVar     | testenvvar
 40 |       4 | hbase          | hbase
 36 |       2 | Scala Test     | scala-test
 46 |       1 | R Project      | r-project
 45 |       4 | spackshell     | spackshell
 37 |       4 | tim            | tim
 48 |      10 | rtest          | rtest
 35 |       1 | Scala Project  | scala-project
 39 |       5 | salim          | salim
 38 |       4 | timtest        | timtest
(14 rows)

To put all above together, I have below quick shell script that can do the job:

for project_id in `ls /var/lib/cdsw/current/projects/projects/0/`
do
  echo "Processing $project_id"
  rows=`kubectl exec $(cdsw status | grep 'db-' | cut -d '|' -f 2 | sed 's/ //g') -ti -- psql -U sense -c "SELECT * FROM projects WHERE ID = $project_id" | grep '0 row' | wc -l`
  if [ $rows -gt 0 ]; then
    echo "Project $project_id has been deleted, you can archive directory /var/lib/cdsw/current/projects/projects/0/$proejct_id"
  fi
done

The output looks like something below:

Processing project 1
Processing project 34
Processing project 35
Processing project 36
Processing project 37
Processing project 38
Processing project 39
Processing project 40
Processing project 41
Processing project 42
Project 42 has been deleted, you can archive directory /var/lib/cdsw/current/projects/projects/0/
Processing project 43
Project 43 has been deleted, you can archive directory /var/lib/cdsw/current/projects/projects/0/
Processing project 44
Processing project 45
Processing project 46
Processing project 47
Processing project 48

So, until there is a fix for this issue, I hope above simple shell script can help.

Any suggestions or ideas, please let me know in the comments section below, thanks a lot in advance.

Leave a Reply

Your email address will not be published. Required fields are marked *