[
https://issues.apache.org/jira/browse/HDDS-7182?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sammi Chen updated HDDS-7182:
-----------------------------
Description:
*max_open_files,* as described in RocksDB document
[https://github.com/facebook/rocksdb/wiki/RocksDB-Tuning-Guide,]
RocksDB keeps all file descriptors in a table cache. If number of file
descriptors exceeds max_open_files, some files are evicted from table cache and
their file descriptors closed. This means that every read must go through the
table cache to lookup the file needed. Set max_open_files to -1 to always keep
all files open, which avoids expensive table cache calls.
The default value is -1 which means on limit. But on OS level, the open file
handlers of each process is actually limited. If there is too many sst files in
RocksDB, there probably will be failure like "Too many open files" in some
extreme cases.
Change the default value from -1 to 1024.
was:
*max_open_files,* as described in RocksDB document
[https://github.com/facebook/rocksdb/wiki/RocksDB-Tuning-Guide,]
RocksDB keeps all file descriptors in a table cache. If number of file
descriptors exceeds max_open_files, some files are evicted from table cache and
their file descriptors closed. This means that every read must go through the
table cache to lookup the file needed. Set max_open_files to -1 to always keep
all files open, which avoids expensive table cache calls.
The default value is -1 which means on limit. But on OS level, the open file
handlers of each process is actually limited. If there is too many sst files in
RocksDB, there probably will be failure like "Too many open files" in some
extreme cases.
Change the default value from -1 to 1024
> Add property to control max open files
> --------------------------------------
>
> Key: HDDS-7182
> URL: https://issues.apache.org/jira/browse/HDDS-7182
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Sammi Chen
> Assignee: Sammi Chen
> Priority: Major
>
> *max_open_files,* as described in RocksDB document
> [https://github.com/facebook/rocksdb/wiki/RocksDB-Tuning-Guide,]
> RocksDB keeps all file descriptors in a table cache. If number of file
> descriptors exceeds max_open_files, some files are evicted from table cache
> and their file descriptors closed. This means that every read must go through
> the table cache to lookup the file needed. Set max_open_files to -1 to always
> keep all files open, which avoids expensive table cache calls.
> The default value is -1 which means on limit. But on OS level, the open file
> handlers of each process is actually limited. If there is too many sst files
> in RocksDB, there probably will be failure like "Too many open files" in
> some extreme cases.
> Change the default value from -1 to 1024.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
