Performance translators, 'io-cache' and 'read-ahead' were not adding any performance improvement. The 'read-ahead' is redundant due to kernel 'read-ahead' and 'io-cache' was degrading the performance for workloads that don’t involve re-read.
As the VFS layer already has both these functionalities, the two translators have been turned off by default for native fuse mounts. Due to this, there is improved performance for the sequential read.
The following steps were performed.
1. No new failures were seen in the automated regression suite after the perf translators io-cache and read-ahead have been turned off.
2. Created all volume types and mounted the volume .
3. Validated the fuse vol and no read-ahead, io-cache translator is loaded.
gluster vol get <volname> all | grep 'read-ahead'
gluster vol get <volname> all | grep 'io-cache'
These translators were disabled .
Hence moving the bug to verified state.
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
https://access.redhat.com/errata/RHBA-2020:0288
The following steps were performed. 1. No new failures were seen in the automated regression suite after the perf translators io-cache and read-ahead have been turned off. 2. Created all volume types and mounted the volume . 3. Validated the fuse vol and no read-ahead, io-cache translator is loaded. gluster vol get <volname> all | grep 'read-ahead' gluster vol get <volname> all | grep 'io-cache' These translators were disabled . Hence moving the bug to verified state.