-
Notifications
You must be signed in to change notification settings - Fork 1.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
User cannot change Shards for some indices (e.g. System, Illuminate) #21254
Comments
Unable to reproduce in master, I am able to change shard config of Graylog Events index |
@tellistone this appears to be mostly an issue with illuminate indices. I did confirm i cannot reproduce this with a greenfield 6.1 environment. I CAN save settings for an illuminate indices. Also is it possible there is something that gets carried over from existing installs vs a greenfield install? How can one work around this issue if they are to encounter it? Is it also possible code changes were made to master that resolved this? This is very important for configuring a Graylog cluster to be HA. |
The error is very strange because it claims i am trying to change the field mappings for illuminate which, not only am i not trying to do, the UX removes the field mapping UI elements and the PUT payload also is not setting any field mappings: {
"title": "Fortinet Event Log Messages",
"description": "Fortinet Event Log Messages",
"index_prefix": "gl_fortinet",
"writable": true,
"can_be_default": true,
"shards": 1,
"data_tiering": {
"type": "hot_only",
"index_lifetime_min": "P14D",
"index_lifetime_max": "P30D"
},
"replicas": 0,
"rotation_strategy_class": "org.graylog2.indexer.rotation.strategies.TimeBasedSizeOptimizingStrategy",
"rotation_strategy": {
"type": "org.graylog2.indexer.rotation.strategies.TimeBasedSizeOptimizingStrategyConfig",
"index_lifetime_min": "P7D",
"index_lifetime_max": "P14D"
},
"retention_strategy_class": "org.graylog2.indexer.retention.strategies.DeletionRetentionStrategy",
"retention_strategy": {
"type": "org.graylog2.indexer.retention.strategies.DeletionRetentionStrategyConfig",
"max_number_of_indices": 7
},
"index_analyzer": "keyword",
"index_optimization_max_num_segments": 1,
"index_optimization_disabled": false,
"field_type_refresh_interval": 30000,
"id": "63c1b02d5707651ef182399a",
"creation_date": "2023-01-13T19:25:33.406Z",
"index_template_type": "illuminate_content",
"use_legacy_rotation": true,
"default": false
} |
So we have a subset of customers who will be unable to modify their illuminate index set config (shard count etc) IF they have any custom mappings created before we locked it down. Would the more appropriate path be to auto-nuke any custom field mappings on these index sets, or to handle the exception of them being in existence? |
I think so. From what I tested previously even though the frontend allowed you to set these on illumiante index sets, they never actually got applied. I checked the old issue Note:
|
Graylog users (anyone using graylog, even admins) are disallowed from modifying the Index Set settings for specific Index Sets.
Attempting to save index set settings for the affected index sets yields the following error:
HTTP 400
See
graylog2-server/graylog2-server/src/main/java/org/graylog2/indexer/IndexSetValidator.java
Line 139 in 6864f1d
Appears to apply to the following types of Index Template Types:
graylog2-server/graylog2-server/src/main/java/org/graylog2/indexer/indexset/IndexSetConfig.java
Line 67 in 6864f1d
It is crucial the user be able to modify at least the shard settings in order to specify the correct number of primary shards and replica shards. Not having the ability to do so can make the OpenSearch cluster brittle and unreliable, causing a Red cluster health and cause stability problems for Graylog, such as the inability to ingest new messages or search existing messages.
Expected Behavior
Users are able to modify crucial shard settings for ALL index sets
Current Behavior
Users are unable to modify the above mentioned index sets
Possible Solution
Allow modification of "Details" section for ALL index sets (Index Shards, Index Replicas, Maximum Number of Segments)
Steps to Reproduce (for bugs)
Via Graylog Web Interface
Graylog Events
)Context
I suffered a fatal error with my 2 node OpenSearch cluster. To avoid future issues I am reconfiguring all index sets to have at least 1 replica shard so that if an OpenSearch node is offline or a shard has a catastrophic write failure, the OpenSearch cluster won't go into a Red Health state preventing log ingestion.
Your Environment
Please let me know if there are any questions.
The text was updated successfully, but these errors were encountered: