starlingx/jenkins-pipelines
Code Issues Proposed changes

publish_utils.sh: faster checksum matching

Browse Source 
There are a large number of checksum files when publishing
aptly.  We need to search them an parallel for checksum matches.
Doing so speeds publication by an order of magnitude.

Signed-off-by: Scott Little <scott.little@windriver.com>
Change-Id: Iedb392e1bd1d63c4755e1b5c8eab8eb8a379bb14
This commit is contained in:
Scott Little 2022-10-28 14:08:29 -04:00
parent 0d1a482f85
commit 7f604733e5
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
scripts/lib/publish_utils.sh
View File

@ -24,12 +24,15 @@ publish_file() {
local checksum
checksum=$(sha256sum "$filename" | awk '{print $1}' ; check_pipe_status) || exit 1
local p
p=$(nproc)
# find an existing file in $combined_checksums_file that we can
# hardlink to
local link_created
link_created=$(
cat "$published_checksum_files_list_file" | {
cat "$published_checksum_files_list_file" | xargs -P $p \grep -l "$checksum " | {
while read checksum_file ; do
local checksum_dir
checksum_dir="${checksum_file%/*}"