#
|
#
|
# The contents of this file are subject to the terms of the Common Development and
|
# Distribution License (the License). You may not use this file except in compliance with the
|
# License.
|
#
|
# You can obtain a copy of the License at legal/CDDLv1.0.txt. See the License for the
|
# specific language governing permission and limitations under the License.
|
#
|
# When distributing Covered Software, include this CDDL Header Notice in each file and include
|
# the License file at legal/CDDLv1.0.txt. If applicable, add the following below the CDDL
|
# Header, with the fields enclosed by brackets [] replaced by your own identifying
|
# information: "Portions copyright [year] [name of copyright owner]".
|
#
|
# Copyright 2026 3A Systems, LLC.
|
|
# The disturbance restart-ds of the replication convergence run (#1208), sourced by run.sh: a
|
# directory server restarts DISTURB_AT seconds into the load, as docker restart or a rolling
|
# update restarts it, while the others go on taking the load. run_load stops its clients
|
# before and starts others once it serves again. It is the second directory server, the one
|
# inject_divergence writes to: in rs534 it runs no replication server, in combined the one it
|
# runs restarts with it.
|
|
DISTURBED=$(echo $DS_NODES | cut -d' ' -f2)
|
|
# docker restart sends SIGTERM and kills the server once its timeout is over: 10 s by default,
|
# less than a server under load may need to stop, and a killed server is another disturbance.
|
# stop_all gives it 120 s as well. The restart goes to disturbances.txt for the summary, from
|
# the restart until the server serves again, and is a window of the errors logs: the handshakes
|
# its start fails on a loaded runner are expected there (replication_errors). The window closes
|
# 10 s after the server serves again: a handshake its start began can still fail then, and a
|
# peer logs 145 only once its wait for the domain lock, of up to 8 s, is over
|
# (ServerHandler.lockDomainWithTimeout). A handshake that fails later is an error again
|
disturb() {
|
local from served
|
echo " [$SECONDS s] restarting $DISTURBED"
|
window_open
|
from=$(date +%s)
|
docker restart -t 120 $DISTURBED >/dev/null
|
wait_until $DISTURB_TIMEOUT "$DISTURBED serves again" started_since $DISTURBED $from
|
served=$(date +%s)
|
echo " [$SECONDS s] $DISTURBED serves again"
|
sleep 10
|
window_close
|
echo "$((from * 1000)) $((served * 1000)) $DISTURBED restart" >>"$OUT/disturbances.txt"
|
}
|
|
# After the load, before the wait for convergence: the join of the restarted server, which runs
|
# again on every start, has ended, as topology_up waits for it. Its container is not waited to
|
# be healthy again: the probe may not get through while the replicas catch up either
|
settle() {
|
wait_until $DISTURB_TIMEOUT "the join of $DISTURBED ends" bootstrap_ended $DISTURBED
|
}
|
