New arkcase-tika pod by cherepnalkovski · Pull Request #84 · ArkCase/ark_helm_charts

cherepnalkovski · 2026-02-06T13:07:33Z

No description provided.

drivera-armedia · 2026-02-06T14:15:56Z

doc/README.md

 - [Pentaho](https://www.hitachivantara.com/en-us/products/dataops-software/data-integration-analytics.html) (for reporting services)
 - [Alfresco](https://www.alfresco.com/), for content storage services
 - [Minio](https://min.io/), for content storage services in S3-compatible mode
+- [Tika](https://tika.apache.org/), for document metadata and text extraction services


Please fix the indentation to maintain consistency

looks like Idea is messing this up, I will fix this trough the UI

drivera-armedia · 2026-02-06T14:16:34Z

src/app/charts/tika/files/config/txt/tika-config.xml

Non-executable files don't require permissions being specified. That extension was added to the scripts specifically for executable files.

drivera-armedia · 2026-02-06T14:17:15Z

src/app/charts/tika/files/config/txt/tika-config.xml

+        <metadataFilter class="com.armedia.acm.tika.filter.ContentTypeNormalizationFilter">
+            <contentTypeFixes>audio/vnd.wave:audio/wav,audio/mpeg:audio/mp3,audio/x-flac:audio/flac,video/quicktime:video/mp4</contentTypeFixes>
+            <extensionOverrides>.mpga:.mp3,.qt:.mp4</extensionOverrides>
+        </metadataFilter>
+        <metadataFilter class="com.armedia.acm.tika.filter.CreatedDateNormalizationFilter"/>
+        <metadataFilter class="com.armedia.acm.tika.filter.GpsEnrichmentFilter"/>


I see com.armedia.acm.* ... do we have to add JARs of our own to extend the Tika container?

Yes, this is how we extend Tika Server capabilities without modifying Tika’s source code.
We provide our custom parser implementations as a separate JAR and load them through the container’s classpath, and then activate/configure them via tika-config.xml.
This follows Tika’s recommended extension mechanism.

Then we need to add those JAR's Nexus/Maven coordinates to the container's build process. Can you provide the list for them or create an MR for the container?

It's already created, but I need the arkcase-tika merged, until then, the jar is not available on nexus.
ArkCase/ark_tika#1

You can still add the coordinates to the Dockerfile, commented out ... same with the RUN commands to fetch them. The point here is to start getting everything lined up.

Furthermore, it might actually behoove us to have those JARs be separate from ArkCase since they're not specifically dependent on ArkCase to begin with ... are they?

Yes, it’s a separate small module outside of ArkCase that contains only the custom parsing logic and filters.

I have tested this locally using the updated ark_tika Docker image (including the arkcase-tika.jar) along with the Helm changes and the new Tika pod. The pod is up and running, the configuration is being picked up correctly, and all custom parsers/filters are properly registered.

At the moment, I’m experiencing issues running the core pod with the latest develop branch. Once I get it running, I’ll be able to fully confirm that the ArkCase–Tika integration is working as expected.

drivera-armedia · 2026-02-06T14:18:14Z

src/app/charts/tika/templates/ark-tika.yaml

+  strategy:
+    type: {{ $.Values.updateStrategy }}
+    {{- if (eq $.Values.updateStrategy "RollingUpdate") }}
+    rollingUpdate:
+      maxSurge: 1
+      maxUnavailable: 0
+    {{- end }}


This can stay as-is for now, but we'll probably want to use a HorizontalPodAutoscaler for this Deployment

On second thought, let's just replace this with:

strategy: type: Recreate

We don't really need to do a rolling update since these are meant to be stateless.

drivera-armedia · 2026-02-06T14:19:38Z

src/app/charts/tika/templates/ark-tika.yaml

+            - name: JAVA_OPTS
+              value: >-
+                -Xmx1536m
+                -Xms512m
+                -XX:+UseG1GC


We should be able to use the percentage-based stuff so we can then control how much memory the pod is allocated via its resource allocations ... look for what's done for other pods like messaging or core

drivera-armedia · 2026-02-06T14:23:03Z

src/app/charts/tika/Chart.yaml

@@ -0,0 +1,10 @@
+apiVersion: v2
+name: tika
+version: 0.10.0-2


The version can remain as 0.10.0-0 since it's the first release of this subchart.

drivera-armedia · 2026-02-06T14:24:18Z

src/app/charts/tika/values.yaml

+persistence:
+  enabled: true
+  volumeSize:
+    init: "1Gi"
+    data: "16Gi"
+  volumes:
+    data:
+      capacity: "16Gi"
+      accessModes:
+        - ReadWriteOnce


I don't think Tika requires persistence, so this can be removed or substituted with:

persistence: enabled: false

drivera-armedia · 2026-02-06T14:25:09Z

src/app/charts/tika/values.yaml

+      httpGet:
+        scheme: HTTP
+        path: "/tika"
+        port: *tika


The scheme should be HTTPS, b/c we should be using SSL everywhere.

Looks like Tika Server 3.x does not support native HTTPS. I tried configuring it via both CLI flags and tika-config.xml, but the server doesn't recognize any SSL/keystore parameters.
How do we handle services that only support HTTP? Can you help with this one ?

I see they have just committed a change to improve the TLS docs: https://www.mail-archive.com/commits@tika.apache.org/msg20561.html

Nice catch, but the fix is planned for version 4.0.0 😢
It’s estimated to be released in January, however it’s still not available and there haven’t been any updates so far.
For now, we’ll probably need to use a proxy as a workaround.

https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=148639291#TikaServer-SSL(Beta)

This is already available since Tika 2.4.0. I'll test this and figure this out. Most likely the challenge had to do with access to the keystore(s).

@drivera-armedia did you had a chance to look at it?

drivera-armedia · 2026-02-06T14:26:34Z

src/app/templates/ark-service-tika.yaml

This second service declaration is unnecessary. The reason some pods have two service declarations is b/c they need to be aware of their cluster members. Tika is not a cluster per-se, and thus doesn't need to know how many other Tika pods are available.

Thus, the above service declaration is sufficient, and this one is unnecessary.

drivera-armedia · 2026-02-06T14:26:59Z

README.md

 - [PostgreSQL](https://www.postgresql.org/)/[MariaDB](https://mariadb.org/), for database storage (only one instance is needed)
 - [Pentaho](https://www.hitachivantara.com/en-us/products/dataops-software/data-integration-analytics.html) (for reporting services)
 - [Alfresco](https://www.alfresco.com/), for content storage services
+- [Tika](https://tika.apache.org/), for document metadata and text extraction services


As above, please fix the indentation to maintain consistency.

jenkins-armedia · 2026-02-09T15:20:58Z

src/app/charts/tika/files/config/txt/tika-config.xml

+        <metadataFilter class="com.armedia.acm.tika.filter.ContentTypeNormalizationFilter">
+            <contentTypeFixes>audio/vnd.wave:audio/wav,audio/mpeg:audio/mp3,audio/x-flac:audio/flac,video/quicktime:video/mp4</contentTypeFixes>
+            <extensionOverrides>.mpga:.mp3,.qt:.mp4</extensionOverrides>
+        </metadataFilter>
+        <metadataFilter class="com.armedia.acm.tika.filter.CreatedDateNormalizationFilter"/>
+        <metadataFilter class="com.armedia.acm.tika.filter.GpsEnrichmentFilter"/>


Then we need to add those JAR's Nexus/Maven coordinates to the container's build process. Can you provide the list for them or create an MR for the container?

jenkins-armedia · 2026-02-09T15:21:35Z

src/app/charts/tika/templates/_helpers.tpl

If there are no specific helpers needed, let's remove this file.

jenkins-armedia · 2026-02-09T15:26:35Z

src/app/charts/tika/templates/ark-tika.yaml

+  strategy:
+    type: {{ $.Values.updateStrategy }}
+    {{- if (eq $.Values.updateStrategy "RollingUpdate") }}
+    rollingUpdate:
+      maxSurge: 1
+      maxUnavailable: 0
+    {{- end }}


On second thought, let's just replace this with:

strategy: type: Recreate

We don't really need to do a rolling update since these are meant to be stateless.

jenkins-armedia · 2026-02-09T15:27:36Z

src/app/charts/tika/templates/ark-tika.yaml

+            - name: INIT_DIR
+              value: &initDir "/app/init"


Why do we need an init volume? This is meant to be a stateless pod - what initialization does it need that can't be done as part of the container's boot sequence?

Actually, we don't need it. It's removed, and DATA_DIR is replaced wiht TEMP_DIR

jenkins-armedia · 2026-02-09T15:27:48Z

src/app/charts/tika/templates/ark-tika.yaml

+                  - ownership: {{ coalesce ($.Values.persistence).ownership "1999:1999" | quote }}
+                    permissions: "u=rwX,g=rX,o="
+                    flags: [ "recurse", "forced", "create", "changes" ]
+                    targets: [ "$(DATA_DIR)", "$(INIT_DIR)" ]


See above re: init volume

jenkins-armedia · 2026-02-09T15:27:52Z

src/app/charts/tika/templates/ark-tika.yaml

+            - name: &initVol "init"
+              mountPath: *initDir


See above re: init volume

jenkins-armedia · 2026-02-09T15:27:57Z

src/app/charts/tika/templates/ark-tika.yaml

+            - name: *initVol
+              mountPath: *initDir


See above re: init volume

jenkins-armedia · 2026-02-09T15:28:08Z

src/app/charts/tika/templates/ark-tika.yaml

+        - name: *initVol
+          emptyDir:
+            sizeLimit: 1Gi


See above re: init volume

… docker image.

drivera-armedia · 2026-02-09T22:20:49Z

src/app/charts/tika/files/config/txt/tika-config.xml

+        <metadataFilter class="com.armedia.acm.tika.filter.ContentTypeNormalizationFilter">
+            <contentTypeFixes>audio/vnd.wave:audio/wav,audio/mpeg:audio/mp3,audio/x-flac:audio/flac,video/quicktime:video/mp4</contentTypeFixes>
+            <extensionOverrides>.mpga:.mp3,.qt:.mp4</extensionOverrides>
+        </metadataFilter>
+        <metadataFilter class="com.armedia.acm.tika.filter.CreatedDateNormalizationFilter"/>
+        <metadataFilter class="com.armedia.acm.tika.filter.GpsEnrichmentFilter"/>


You can still add the coordinates to the Dockerfile, commented out ... same with the RUN commands to fetch them. The point here is to start getting everything lined up.

Furthermore, it might actually behoove us to have those JARs be separate from ArkCase since they're not specifically dependent on ArkCase to begin with ... are they?

drivera-armedia · 2026-02-09T22:21:37Z

src/app/charts/tika/values.yaml

+      httpGet:
+        scheme: HTTP
+        path: "/tika"
+        port: *tika


https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=148639291#TikaServer-SSL(Beta)

This is already available since Tika 2.4.0. I'll test this and figure this out. Most likely the challenge had to do with access to the keystore(s).

drivera-armedia · 2026-02-10T16:01:50Z

src/app/charts/tika/templates/service.yaml

@@ -0,0 +1,3 @@
+{{- if not (include "arkcase.subsystem.external" $) -}}
+  {{- include "arkcase.subsystem.service" (dict "ctx" $ "dnsOnly" true) -}}


Please address this one

cherepnalkovski changed the base branch from main to develop February 6, 2026 13:12

cherepnalkovski force-pushed the tika-setup branch from 186364b to 9a9c5f5 Compare February 6, 2026 13:15

Setup new arkcase-tika pod

f473b8c

cherepnalkovski force-pushed the tika-setup branch from 9a9c5f5 to f473b8c Compare February 6, 2026 13:52

drivera-armedia requested changes Feb 6, 2026

View reviewed changes

Vladimir Cherepnalkovski and others added 4 commits February 8, 2026 20:16

set persistence false, remove pvc, fix formating and version

e885d00

Fix formatting of Tika entry in README

e3f93e7

Fix formatting for Tika entry in README

2eb58ce

remove inherited java options

43aaf40

jenkins-armedia requested changes Feb 9, 2026

View reviewed changes

Vladimir Cherepnalkovski added 3 commits February 9, 2026 20:24

remove not needed file

f4ce5b9

Change from rollingUpgrade to recreate strategy

22201da

Remove the init volume and rename data to temp as it's defined in the…

e08dd2a

… docker image.

ArkCase deleted a comment from jenkins-armedia Feb 9, 2026

drivera-armedia requested changes Feb 9, 2026

View reviewed changes

cherepnalkovski requested a review from drivera-armedia February 10, 2026 11:58

drivera-armedia requested changes Feb 10, 2026

View reviewed changes

Vladimir Cherepnalkovski added 3 commits February 11, 2026 00:24

remove the dnsOnly flag.

249e9c9

Remove hardcoded image ark_tika is already available

8219ee3

Merge branch 'develop' into tika-setup

bdf5bfa

cherepnalkovski requested review from david-oc-miller, drivera-armedia and jenkins-armedia February 17, 2026 11:00

drivera-armedia added 4 commits February 17, 2026 08:02

Merge branch 'develop' into tika-setup

e59436a

Merge branch 'develop' into tika-setup

dd21319

Add <server> section

d40e4af

Make sure the keystore is also a truststore

2b8a143

		@@ -0,0 +1,3 @@
		{{- if not (include "arkcase.subsystem.external" $) -}}
		{{- include "arkcase.subsystem.service" (dict "ctx" $ "dnsOnly" true) -}}

Conversation

cherepnalkovski commented Feb 6, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cherepnalkovski Feb 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

cherepnalkovski Feb 17, 2026 •

edited

Loading