Releases: dnanexus/dxfuse
v1.4.0 Fix truncated reads from cache
v1.3.0 List folder objects and describe content in a single request
List folder objects and describe content in a single /listFolder
request instead of /listFolder
and then iterating with /system/findDataObjects
.
Update MacOS runner version in GHA.
v1.2.0 Fix dnanexus files that are symlinks
Fix downloading dnanexus files that are backed by symlinks. Previously this only worked for symlinks on a public s3 bucket.
Update Linux build from Ubuntu 18.04 --> 20.04 github action runner as 18.04 was deprecated.
v1.1.1 Remove command server
Remove the command server functionality that was previously used in dxfuse -sync
which was deprecated in v1.0.0
. This will allow multiple dxfuse processes to run on the same machine without port conflicts.
v1.1.0 Write temporary manifest to $HOME/.dxfuse/
Write temporary manifest to $HOME/.dxfuse/
instead of /tmp/
to allow multiple users to run an instance dxfuse concurrently on the same machine.
v1.0.1 Fix crash when file-xxxx/download returns error
Switch back to go-sqlite3 with CGO_ENABLED build for sqlite dependency.
Fix crash when file-xxxx/download returns error
v1.0.0 -limitedWrite mode
Backwards incompatible release v1.0.0.
New flag -limitedWrite for supporting append only writes to DNAnexus files. See README for supported operations and limitations of this mode.
Remove dxfuse -sync command previously used for syncing writeable files to DNAnexus.
Enable FUSE options for kernel readahead and pagecache. Enabling pagecache adds support for shared read-only mmap which is used by many applications, e.g. gatk
, tidyverse readr
. Kernel readahead fixes some dxfuse crashes with prefetching and may be more stable.
v1.0.0-rc.5 Fix RmDir of spark temporary output folders
Fix filepaths to use filepath
package instead of concatenating strings with /
Bump golang version to 1.16.10
v1.0.0-rc.4 Raise min http client pool
Increase minimum http client pool to 30. Lower wait time to check if file has been closed to 400ms.
v1.0.0-rc.3 Limit concurrent file writes
Attempt to reduce memory usage by limiting the number of files which can be written to concurrently. The limit is greater of 15 or CPUs*3
Lower the initial upload file part size from 96MiB --> 16MiB. Each subsequent part n
increases in size 16 * 1.1^n
MiB up to a maximum part size of 700MiB.
Raise the HttpClient pool used for DNAnexus api calls to a minimum of 15 and up to CPUs*3
.