From aa5c38727b314b03cd7ab69612435aa206bd5e2c Mon Sep 17 00:00:00 2001
From: Dan Smith <dansmith@redhat.com>
Date: Wed, 14 Apr 2021 14:27:32 -0700
Subject: [PATCH] Work around CHILD_MAX bash limitation for async

Apparently bash (via POSIX) only guarantees a small (32ish) number of
children can be started and their statuses retrieved at any given
point. On larger jobs with lots of plugins and additional work, we
may go over that limit, especially for long-lived children, such
as the install_tempest task.

This works around that issue by creating a fifo for each child at
spawn time. When the child is complete, it will block on a read
against that fifo (and thus not exit). When the parent goes to wait
on the child, it first writes to that fifo, unblocking the child so
that it can exit near the time we go to wait.

Closes-Bug: #1923728
Change-Id: Id755bdb1e7f1664ec08742d034c174e87a3d2902
---
 inc/async | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/inc/async b/inc/async
index c63bc2045a..11bcdfa39e 100644
--- a/inc/async
+++ b/inc/async
@@ -57,6 +57,7 @@ function async_log {
 function async_inner {
     local name="$1"
     local rc
+    local fifo=${DEST}/async/${name}.fifo
     shift
     set -o xtrace
     if $* >${DEST}/async/${name}.log 2>&1; then
@@ -69,6 +70,8 @@ function async_inner {
         async_log "$name" "FAILED with rc $rc"
     fi
     iniset ${DEST}/async/${name}.ini job end_time $(date "+%s%3N")
+    # Block on the fifo until we are signaled to exit by the main process
+    cat $fifo
     return $rc
 }
 
@@ -86,12 +89,14 @@ function async_run {
     local name="$1"
     shift
     local inifile=${DEST}/async/${name}.ini
+    local fifo=${DEST}/async/${name}.fifo
 
     touch $inifile
     iniset $inifile job command "$*"
     iniset $inifile job start_time $(date +%s%3N)
 
     if [[ "$DEVSTACK_PARALLEL" = "True" ]]; then
+        mkfifo $fifo
         async_inner $name $* &
         iniset $inifile job pid $!
         async_log "$name" "running: %command"
@@ -119,17 +124,23 @@ function async_wait {
     xtrace=$(set +o | grep xtrace)
     set +o xtrace
 
-    local pid rc running inifile runtime
+    local pid rc running inifile runtime fifo
     rc=0
     for name in $*; do
         running=$(ls ${DEST}/async/*.ini 2>/dev/null | wc -l)
         inifile="${DEST}/async/${name}.ini"
+        fifo=${DEST}/async/${name}.fifo
 
         if pid=$(async_pidof "$name"); then
             async_log "$name" "Waiting for completion of %command" \
                       "($running other jobs running)"
             time_start async_wait
             if [[ "$pid" != "self" ]]; then
+                # Signal the child to go ahead and exit since we are about to
+                # wait for it to collect its status.
+                echo "Signaling exit"
+                echo WAKEUP > $fifo
+                echo "Signaled"
                 # Do not actually call wait if we ran synchronously
                 if wait $pid; then
                     rc=0
@@ -137,6 +148,7 @@ function async_wait {
                     rc=$?
                 fi
                 cat ${DEST}/async/${name}.log
+                rm -f $fifo
             fi
             time_stop async_wait
             local start_time