Set coredump filter for Elasticsearch #137300
Merged
+75
−2
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
When a linux process dies and the core process memory is dumped, which parts of the memory to output are controlled by the coredump filter. By default the JVM dumps anonymous memory (ie heap and direct memory) and file-backed memory (ie mmaps), and private huge pages. However, the default of dumping file-backed memory for Elasticsearch is problematic; these can be enormous, and they also aren't very useful since the indices will likely be available when investigating the crash. This commit sets the coredump filter when running on Linux to omit file-backed memory from coredumps. Although the JVM has diagnostic VM options for controlling the filter, they end up causing the filter to be inherited, which is from the server CLI, another Java process, which has the file-backed memory bits set. Instead of using the JVM options we set it via the special proc filesystem.
rjernst
added
>refactoring
:Core/Infra/Node Lifecycle
elasticsearchmachine
added
v9.3.0
Team:Core/Infra
|
Pinging @elastic/es-core-infra (Team:Core/Infra) |
|
Note for reviewers: there isn't any easy way to test this. Ideally we could have a packaging test, but the process id we have there is the CLI. So we'll need to look at children of the CLI... |
ldematte
approved these changes
Oct 29, 2025
| // which is the server CLI, which is also a JVM so it has these bits set. Thus, we set it explicitly. | ||
| // See https://man7.org/linux/man-pages/man5/core.5.html for more info on the relevant bits of the filter | ||
| try { | ||
| Files.writeString(Paths.get("/proc/self/coredump_filter"), "0x23"); |
There was a problem hiding this comment.
I think this either needs a @SuppressForbidden or using and alternative (PathUtils?)
This was referenced Oct 31, 2025
rjernst
added a commit
to rjernst/elasticsearch
that referenced
this pull request
Oct 31, 2025
When a linux process dies and the core process memory is dumped, which parts of the memory to output are controlled by the coredump filter. By default the JVM dumps anonymous memory (ie heap and direct memory) and file-backed memory (ie mmaps), and private huge pages. However, the default of dumping file-backed memory for Elasticsearch is problematic; these can be enormous, and they also aren't very useful since the indices will likely be available when investigating the crash. This commit sets the coredump filter when running on Linux to omit file-backed memory from coredumps. Although the JVM has diagnostic VM options for controlling the filter, they end up causing the filter to be inherited, which is from the server CLI, another Java process, which has the file-backed memory bits set. Instead of using the JVM options we set it via the special proc filesystem.
rjernst
added a commit
to rjernst/elasticsearch
that referenced
this pull request
Oct 31, 2025
When a linux process dies and the core process memory is dumped, which parts of the memory to output are controlled by the coredump filter. By default the JVM dumps anonymous memory (ie heap and direct memory) and file-backed memory (ie mmaps), and private huge pages. However, the default of dumping file-backed memory for Elasticsearch is problematic; these can be enormous, and they also aren't very useful since the indices will likely be available when investigating the crash. This commit sets the coredump filter when running on Linux to omit file-backed memory from coredumps. Although the JVM has diagnostic VM options for controlling the filter, they end up causing the filter to be inherited, which is from the server CLI, another Java process, which has the file-backed memory bits set. Instead of using the JVM options we set it via the special proc filesystem.
rjernst
added a commit
to rjernst/elasticsearch
that referenced
this pull request
Oct 31, 2025
When a linux process dies and the core process memory is dumped, which parts of the memory to output are controlled by the coredump filter. By default the JVM dumps anonymous memory (ie heap and direct memory) and file-backed memory (ie mmaps), and private huge pages. However, the default of dumping file-backed memory for Elasticsearch is problematic; these can be enormous, and they also aren't very useful since the indices will likely be available when investigating the crash. This commit sets the coredump filter when running on Linux to omit file-backed memory from coredumps. Although the JVM has diagnostic VM options for controlling the filter, they end up causing the filter to be inherited, which is from the server CLI, another Java process, which has the file-backed memory bits set. Instead of using the JVM options we set it via the special proc filesystem.
elasticsearchmachine
pushed a commit
that referenced
this pull request
Oct 31, 2025
When a linux process dies and the core process memory is dumped, which parts of the memory to output are controlled by the coredump filter. By default the JVM dumps anonymous memory (ie heap and direct memory) and file-backed memory (ie mmaps), and private huge pages. However, the default of dumping file-backed memory for Elasticsearch is problematic; these can be enormous, and they also aren't very useful since the indices will likely be available when investigating the crash. This commit sets the coredump filter when running on Linux to omit file-backed memory from coredumps. Although the JVM has diagnostic VM options for controlling the filter, they end up causing the filter to be inherited, which is from the server CLI, another Java process, which has the file-backed memory bits set. Instead of using the JVM options we set it via the special proc filesystem.
elasticsearchmachine
pushed a commit
that referenced
this pull request
Oct 31, 2025
When a linux process dies and the core process memory is dumped, which parts of the memory to output are controlled by the coredump filter. By default the JVM dumps anonymous memory (ie heap and direct memory) and file-backed memory (ie mmaps), and private huge pages. However, the default of dumping file-backed memory for Elasticsearch is problematic; these can be enormous, and they also aren't very useful since the indices will likely be available when investigating the crash. This commit sets the coredump filter when running on Linux to omit file-backed memory from coredumps. Although the JVM has diagnostic VM options for controlling the filter, they end up causing the filter to be inherited, which is from the server CLI, another Java process, which has the file-backed memory bits set. Instead of using the JVM options we set it via the special proc filesystem.
elasticsearchmachine
pushed a commit
that referenced
this pull request
Oct 31, 2025
When a linux process dies and the core process memory is dumped, which parts of the memory to output are controlled by the coredump filter. By default the JVM dumps anonymous memory (ie heap and direct memory) and file-backed memory (ie mmaps), and private huge pages. However, the default of dumping file-backed memory for Elasticsearch is problematic; these can be enormous, and they also aren't very useful since the indices will likely be available when investigating the crash. This commit sets the coredump filter when running on Linux to omit file-backed memory from coredumps. Although the JVM has diagnostic VM options for controlling the filter, they end up causing the filter to be inherited, which is from the server CLI, another Java process, which has the file-backed memory bits set. Instead of using the JVM options we set it via the special proc filesystem.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When a linux process dies and the core process memory is dumped, which parts of the memory to output are controlled by the coredump filter. By default the JVM dumps anonymous memory (ie heap and direct memory) and file-backed memory (ie mmaps), and private huge pages. However, the default of dumping file-backed memory for Elasticsearch is problematic; these can be enormous, and they also aren't very useful since the indices will likely be available when investigating the crash.
This commit sets the coredump filter when running on Linux to omit file-backed memory from coredumps. Although the JVM has diagnostic VM options for controlling the filter, they end up causing the filter to be inherited, which is from the server CLI, another Java process, which has the file-backed memory bits set. Instead of using the JVM options we set it via the special proc filesystem.