Use JNA to Speed up Snapshot Cache File Creation #68687
Merged
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
Use JNA to speed up snapshot cache file creation. Do this in `:server` to bypass the security filter and move necessary bits of code to `:server` to enable the logic. Fall back to trying to create the file by writing zeros if anything except for the step determining free disk space fails.
original-brownbear
added
>non-issue
:Distributed Coordination/Snapshot/Restore
elasticmachine
added
the
Team:Distributed (Obsolete)
|
Pinging @elastic/es-distributed (Team:Distributed) |
| @@ -260,4 +269,39 @@ static void tryInstallSystemCallFilter(Path tmpFile) { | |||
| logger.warn("unable to install syscall filter: ", e); | |||
| } | |||
| } | |||
|
|
|||
| @SuppressForbidden(reason = "need access to fd on FileOutputStream") | |||
| static void fallocateSnapshotCacheFile(Environment environment, long fileSize) throws IOException { | |||
There was a problem hiding this comment.
It's a little unfortunate to leak that much logic into this class, but I figured abstracting things away further to isolate the fallocate as much as possible wasn't really worth the extra complication (IMO).
ywelsch
reviewed
Feb 9, 2021
server/src/main/java/org/elasticsearch/snapshots/SnapshotsService.java
Outdated
Show resolved
Hide resolved
server/src/main/java/org/elasticsearch/snapshots/SnapshotsService.java
Outdated
Show resolved
Hide resolved
server/src/main/java/org/elasticsearch/snapshots/SnapshotsService.java
Outdated
Show resolved
Hide resolved
|
@elasticmachine update branch |
ywelsch
approved these changes
Feb 9, 2021
server/src/main/java/org/elasticsearch/bootstrap/JNANatives.java
Outdated
Show resolved
Hide resolved
This was caused by elastic#68709 which turned the edge case of a range with start > end into a tripped assertion.
tlrx
approved these changes
Feb 10, 2021
| */ | ||
| public static void tryCreateCacheFile(Environment environment, long fileSize) throws IOException { | ||
| if (JNA_AVAILABLE == false) { | ||
| logger.warn("cannot use fallocate to create cache file because JNA is not available"); |
There was a problem hiding this comment.
nit: I know we are not consistent in starting log messages with uppercase or not but we're using both ways in this PR
server/src/main/java/org/elasticsearch/snapshots/SnapshotsService.java
Outdated
Show resolved
Hide resolved
|
Thanks Yannick + Tanguy! |
original-brownbear
added a commit
to original-brownbear/elasticsearch
that referenced
this pull request
Feb 10, 2021
Use JNA to speed up snapshot cache file creation. Do this in `:server` to bypass the security filter and move necessary bits of code to `:server` to enable the logic. Fall back to trying to create the file by writing zeros if anything except for the step determining free disk space fails.
original-brownbear
added a commit
that referenced
this pull request
Feb 10, 2021
Use JNA to speed up snapshot cache file creation. Do this in `:server` to bypass the security filter and move necessary bits of code to `:server` to enable the logic. Fall back to trying to create the file by writing zeros if anything except for the step determining free disk space fails.
joegallo
added a commit
that referenced
this pull request
Feb 10, 2021
|
Reverted in 4c2e9a6 because it broke local development on a mac. (Sorry!) |
original-brownbear
added a commit
to original-brownbear/elasticsearch
that referenced
this pull request
Feb 10, 2021
Second go at elastic#68687 now using the proper indirection so that this works on both OSX and Linux. We can't use the same class we use for `mlock` like the original PR did because the method signature differs across OSX and Linux.
original-brownbear
added a commit
that referenced
this pull request
Feb 10, 2021
Second go at #68687 now using the proper indirection so that this works on both OSX and Linux. We can't use the same class we use for `mlock` like the original PR did because the method signature differs across OSX and Linux.
original-brownbear
added a commit
to original-brownbear/elasticsearch
that referenced
this pull request
Feb 11, 2021
Second go at elastic#68687 now using the proper indirection so that this works on both OSX and Linux. We can't use the same class we use for `mlock` like the original PR did because the method signature differs across OSX and Linux.
original-brownbear
added a commit
that referenced
this pull request
Feb 11, 2021
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Use JNA to speed up snapshot cache file creation. Do this in
:serverto bypass the security filter and move necessary bits of code to
:serverto enable the logic.
Fall back to trying to create the file by writing zeros if anything except for
the step determining free disk space fails.
Manually verified via
duthat this works as expected and actually allocates hundreds of GB in a fewms.