dweiss merged PR #11993:
URL: https://github.com/apache/lucene/pull/11993
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apac
dweiss merged PR #11994:
URL: https://github.com/apache/lucene/pull/11994
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apac
rmuir commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335235418
> @uschindler I think this issue just have a GC path of ThreadLocals. BUT,
for instance in ES, when there is a 1000-shard-nodes, and normally one shard
with 40 segments per shard, one open
rmuir commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335237178
In fact there is no situation where thousands of shards makes sense on a
single node. That's bad design.
--
This is an automated message from the Apache Git Service.
To respond to the me
benwtrent commented on issue #11963:
URL: https://github.com/apache/lucene/issues/11963#issuecomment-1335239951
@rmuir thank you! I currently have a major part of the refactor already
written locally. But, it will be big and will need its own rounds of review to
make sure its the way we wan
benwtrent commented on code in PR #11860:
URL: https://github.com/apache/lucene/pull/11860#discussion_r1038154379
##
lucene/backward-codecs/src/java/org/apache/lucene/backward_codecs/lucene94/ExpandingVectorValues.java:
##
@@ -0,0 +1,49 @@
+/*
+ * Licensed to the Apache Software
rmuir commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335269730
I will investigate the document API and try to make a proposal so that
threadlocal is no longer needed. I'm really concerned about the merge case here
causing regressions for folks that de
luyuncheng commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335302726
> In fact there is no situation where thousands of shards makes sense on a
single node. That's bad design.
@rmuir i have another proposal:
what do you think of make `ByteBuf
jpountz commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335358751
I had opened a very similar PR to this one at
https://github.com/apache/lucene/pull/137 which handled the merge case.
--
This is an automated message from the Apache Git Service.
To re
jpountz commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335365755
I think I had not merged it because the follow-up discussion about removing
thread locals had triggered naming/API concerns, but it should be a good
incremental step and we could figure
rmuir commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335391991
> I think I had not merged it because the follow-up discussion about
removing thread locals had triggered naming/API concerns, but it should be a
good incremental step and we could figure
jpountz commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335392646
Sounds good.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To uns
jpountz closed pull request #137: LUCENE-9955: Reduced state of stored fields
readers.
URL: https://github.com/apache/lucene/pull/137
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific com
rmuir commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335398330
its just gonna take me some time, i can't get something out there like
today. for example nearly 100% of tests would be impacted :) It is fair, I will
feel the same pain the users will. Bu
luyuncheng commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335420862
Threadlocal just scale up the StoredFieldsReader's heap useage, BUT when one
instance with only 10K segments would use 1G heap memory in
`fieldsReaderOrig#buffer` without threadlocal.
rmuir commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335444854
the idea is not to have one instance per segment. there would be zero
instances. when you want to retrieve docs from indexreader, the user would call
.getFieldsReader() or similar to creat
rmuir commented on PR #11987:
URL: https://github.com/apache/lucene/pull/11987#issuecomment-1335456484
and one idea i have is to try to prototype with the term vectors first
(since both stored fields and term vectors have per-segment threadlocals that
I'd like to remove). it is just less te
dnhatn commented on PR #11984:
URL: https://github.com/apache/lucene/pull/11984#issuecomment-1335560505
Thanks @costin!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To u
dnhatn merged PR #11984:
URL: https://github.com/apache/lucene/pull/11984
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apac
dnhatn closed issue #11676: Can TimeLimitingBulkScorer exponentially grow the
window size? [LUCENE-10640]
URL: https://github.com/apache/lucene/issues/11676
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to
jpountz commented on code in PR #11860:
URL: https://github.com/apache/lucene/pull/11860#discussion_r1038244235
##
lucene/backward-codecs/src/java/org/apache/lucene/backward_codecs/lucene94/ExpandingVectorValues.java:
##
@@ -0,0 +1,49 @@
+/*
+ * Licensed to the Apache Software F
benwtrent commented on code in PR #11860:
URL: https://github.com/apache/lucene/pull/11860#discussion_r1038415190
##
lucene/core/src/java/org/apache/lucene/codecs/lucene95/Lucene95HnswVectorsWriter.java:
##
@@ -0,0 +1,753 @@
+/*
+ * Licensed to the Apache Software Foundation (AS
benwtrent commented on code in PR #11860:
URL: https://github.com/apache/lucene/pull/11860#discussion_r1038420089
##
lucene/core/src/java/org/apache/lucene/util/hnsw/HnswGraph.java:
##
@@ -144,6 +145,10 @@ public NodesIterator(int size) {
this.size = size;
}
+p
23 matches
Mail list logo
