Refactor graceful manager, fix misused WaitGroup (#29738)
Follow #29629 (cherry picked from commit d08f4360c96e130e0454b76ecef9405f2bd312a1)
This commit is contained in:
parent
1a17c39e00
commit
94c70c7753
4 changed files with 55 additions and 51 deletions
|
@ -57,20 +57,27 @@ func (g *Manager) start() {
|
|||
// Handle clean up of unused provided listeners and delayed start-up
|
||||
startupDone := make(chan struct{})
|
||||
go func() {
|
||||
defer close(startupDone)
|
||||
// Wait till we're done getting all the listeners and then close the unused ones
|
||||
func() {
|
||||
// FIXME: there is a fundamental design problem of the "manager" and the "wait group".
|
||||
// If nothing has started, the "Wait" just panics: sync: WaitGroup is reused before previous Wait has returned
|
||||
// There is no clear solution besides a complete rewriting of the "manager"
|
||||
defer func() {
|
||||
_ = recover()
|
||||
}()
|
||||
g.createServerWaitGroup.Wait()
|
||||
defer func() {
|
||||
close(startupDone)
|
||||
// Close the unused listeners and ignore the error here there's not much we can do with it, they're logged in the CloseProvidedListeners function
|
||||
_ = CloseProvidedListeners()
|
||||
}()
|
||||
// Ignore the error here there's not much we can do with it, they're logged in the CloseProvidedListeners function
|
||||
_ = CloseProvidedListeners()
|
||||
g.notify(readyMsg)
|
||||
// Wait for all servers to be created
|
||||
g.createServerCond.L.Lock()
|
||||
for {
|
||||
if g.createdServer >= numberOfServersToCreate {
|
||||
g.createServerCond.L.Unlock()
|
||||
g.notify(readyMsg)
|
||||
return
|
||||
}
|
||||
select {
|
||||
case <-g.IsShutdown():
|
||||
g.createServerCond.L.Unlock()
|
||||
return
|
||||
default:
|
||||
}
|
||||
g.createServerCond.Wait()
|
||||
}
|
||||
}()
|
||||
if setting.StartupTimeout > 0 {
|
||||
go func() {
|
||||
|
@ -78,16 +85,7 @@ func (g *Manager) start() {
|
|||
case <-startupDone:
|
||||
return
|
||||
case <-g.IsShutdown():
|
||||
func() {
|
||||
// When WaitGroup counter goes negative it will panic - we don't care about this so we can just ignore it.
|
||||
defer func() {
|
||||
_ = recover()
|
||||
}()
|
||||
// Ensure that the createServerWaitGroup stops waiting
|
||||
for {
|
||||
g.createServerWaitGroup.Done()
|
||||
}
|
||||
}()
|
||||
g.createServerCond.Signal()
|
||||
return
|
||||
case <-time.After(setting.StartupTimeout):
|
||||
log.Error("Startup took too long! Shutting down")
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue