Files
minio/docs/distributed/distributed-from-config-file.sh
T
Feng Ruohang 6613c2a3cb test: pin the external fixtures and run the suites against the silo binary
The test and verification scripts invoked ./minio and pulled their tooling from
upstream infrastructure with no integrity check. Every `curl | tar` of a client
or an old server binary was an unverified execution path in a script that
regularly runs as a privileged user, and several fetched a floating "latest".

Two installers replace all of it:

- install-mcli.sh resolves a pinned pgsty/mc release, downloads the archive and
  its checksum manifest, requires exactly one valid manifest entry for the
  asset, verifies it, and installs. MCLI_BIN with a mandatory MCLI_SHA256 lets
  an offline or air-gapped run supply its own binary, still checksum-checked.
- install-verified-fixture.sh takes source, expected SHA-256 and target, and
  refuses anything that does not match. Sources may be a URL or a local file.

Every script that previously downloaded mc now calls install-mcli.sh. The three
places that genuinely need an upstream artifact - the old MinIO server binary
for the LDAP IAM upgrade-import test, the 2021 mc for the three-site
replication test, and the functional-tests.sh fixture - go through
install-verified-fixture.sh with the digest recorded inline. Those dl.min.io
URLs remain on purpose: they are historical upstream artifacts needed to prove
upgrade compatibility, and they are now pinned and verified rather than
trusted.

The scripts otherwise switch to ./silo, silo.service, the silo container and
compose service names, and SILO_CONFIG_DIR. run-multi-site-minio-idp.sh is
renamed to run-multi-site-silo-idp.sh with the Makefile target following.
buildscripts/minio-upgrade.sh keeps its name and its `minio server` argv - it
exists to test the MinIO-to-Silo upgrade, so the old side must stay old - but
it is now pinned to an image digest rather than a tag, and its `docker system
prune` and `docker volume prune` calls are removed. Those ran unfiltered
against the developer's whole Docker installation; the resiliency tests had the
same problem and lose their prune and `docker ps -q` sweeps too.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-06 08:48:58 +08:00

109 lines
3.2 KiB
Bash
Executable File

#!/usr/bin/env bash
set -e
cleanup() {
echo "Cleaning up instances of Silo"
pkill silo || true
pkill -9 silo || true
rm -rf /tmp/xl/ || true
rm -rf /tmp/minio.configfile.{1,2,3,4} || true
}
cleanup
unset MINIO_KMS_KES_CERT_FILE
unset MINIO_KMS_KES_KEY_FILE
unset MINIO_KMS_KES_ENDPOINT
unset MINIO_KMS_KES_KEY_NAME
export MINIO_CI_CD=1
if [ ! -f ./mc ]; then
os="$(uname -s)"
arch="$(uname -m)"
case "${arch}" in
"x86_64")
arch="amd64"
;;
esac
"$(git rev-parse --show-toplevel)/buildscripts/install-mcli.sh" ./mc
fi
for i in $(seq 1 4); do
s3Port="$((9000 + i))"
consolePort="$((s3Port + 1000))"
cat <<EOF >/tmp/minio.configfile.$i
version: v1
address: ':${s3Port}'
console-address: ':${consolePort}'
rootUser: 'minr0otUS2r'
rootPassword: 'pBU94AGAY85e'
pools: # Specify the nodes and drives with pools
-
- 'http://localhost:9001/tmp/xl/node9001/mnt/disk{1...4}/'
- 'http://localhost:9002/tmp/xl/node9002/mnt/disk{1,2,3,4}/'
-
- 'http://localhost:9003/tmp/xl/node9003/mnt/disk{1...4}/'
- 'http://localhost:9004/tmp/xl/node9004/mnt/disk1/'
- 'http://localhost:9004/tmp/xl/node9004/mnt/disk2/'
- 'http://localhost:9004/tmp/xl/node9004/mnt/disk3/'
- 'http://localhost:9004/tmp/xl/node9004/mnt/disk4/'
EOF
done
silo server --config /tmp/minio.configfile.1 >/tmp/silo1_1.log 2>&1 &
site1_pid=$!
silo server --config /tmp/minio.configfile.2 >/tmp/silo2_1.log 2>&1 &
site2_pid=$!
silo server --config /tmp/minio.configfile.3 >/tmp/silo3_1.log 2>&1 &
site3_pid=$!
silo server --config /tmp/minio.configfile.4 >/tmp/silo4_1.log 2>&1 &
site4_pid=$!
export MC_HOST_silo1=http://minr0otUS2r:pBU94AGAY85e@localhost:9001
export MC_HOST_silo3=http://minr0otUS2r:pBU94AGAY85e@localhost:9003
./mc ready silo1
./mc ready silo3
./mc mb silo1/testbucket
# copy large upload to newbucket on silo1
truncate -s 17M lrgfile
expected_checksum=$(cat ./lrgfile | md5sum)
./mc cp ./lrgfile silo1/testbucket
actual_checksum=$(./mc cat silo3/testbucket/lrgfile | md5sum)
if [ "${expected_checksum}" != "${actual_checksum}" ]; then
echo "unexpected object checksum, expected: ${expected_checksum} got: ${actual_checksum}"
exit
fi
# Compare the difference of the list of disks and their location, with the below expected output
diff <(./mc admin info silo1 --json | jq -r '.info.servers[].drives[] | "\(.pool_index),\(.set_index),\(.disk_index) \(.endpoint)"' | sort) <(
cat <<EOF
0,0,0 http://localhost:9001/tmp/xl/node9001/mnt/disk1
0,0,1 http://localhost:9002/tmp/xl/node9002/mnt/disk1
0,0,2 http://localhost:9001/tmp/xl/node9001/mnt/disk2
0,0,3 http://localhost:9002/tmp/xl/node9002/mnt/disk2
0,0,4 http://localhost:9001/tmp/xl/node9001/mnt/disk3
0,0,5 http://localhost:9002/tmp/xl/node9002/mnt/disk3
0,0,6 http://localhost:9001/tmp/xl/node9001/mnt/disk4
0,0,7 http://localhost:9002/tmp/xl/node9002/mnt/disk4
1,0,0 http://localhost:9003/tmp/xl/node9003/mnt/disk1
1,0,1 http://localhost:9004/tmp/xl/node9004/mnt/disk1
1,0,2 http://localhost:9003/tmp/xl/node9003/mnt/disk2
1,0,3 http://localhost:9004/tmp/xl/node9004/mnt/disk2
1,0,4 http://localhost:9003/tmp/xl/node9003/mnt/disk3
1,0,5 http://localhost:9004/tmp/xl/node9004/mnt/disk3
1,0,6 http://localhost:9003/tmp/xl/node9003/mnt/disk4
1,0,7 http://localhost:9004/tmp/xl/node9004/mnt/disk4
EOF
)
cleanup