Skip to content

fix(generators): own backoff timing in the worker to fix ExponentialBackOff data race (PIPE-1222) #269

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Dylan-M wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix(generators): own backoff timing in the worker to fix ExponentialBackOff data race (PIPE-1222) #269

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 9 additions & 3 deletions generator/apache/apache.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,20 +134,24 @@ func (g *ApacheLogGenerator) worker(workerID int) {
backoffConfig.MaxInterval = 5 * time.Second
backoffConfig.MaxElapsedTime = 0 // Never stop retrying

backoffTicker := backoff.NewTicker(backoffConfig)
defer backoffTicker.Stop()
// Drive the timer from this goroutine only. backoff.ExponentialBackOff is
// not safe for concurrent use, so we never hand it to backoff.NewTicker's
// internal goroutine; instead we own every NextBackOff/Reset call here.
timer := time.NewTimer(backoffConfig.NextBackOff())
defer timer.Stop()

for {
select {
case <-g.stopCh:
g.logger.Debug("Worker stopping", zap.Int("worker_id", workerID))
return
case <-backoffTicker.C:
case <-timer.C:
if g.tracker != nil && !g.tracker.Acquire() {
select {
case <-g.stopCh:
return
case <-g.tracker.ResumeC():
timer.Reset(backoffConfig.NextBackOff())
continue
}
}
Expand All @@ -156,9 +160,11 @@ func (g *ApacheLogGenerator) worker(workerID int) {
g.logger.Error("Failed to write log",
zap.Int("worker_id", workerID),
zap.Error(err))
timer.Reset(backoffConfig.NextBackOff())
continue
}
backoffConfig.Reset()
timer.Reset(backoffConfig.NextBackOff())
}
}
}
Expand Down
12 changes: 9 additions & 3 deletions generator/apache_combined/apache_combined.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -131,20 +131,24 @@ func (g *ApacheCombinedLogGenerator) worker(workerID int) {
backoffConfig.MaxInterval = 5 * time.Second
backoffConfig.MaxElapsedTime = 0 // Never stop retrying

backoffTicker := backoff.NewTicker(backoffConfig)
defer backoffTicker.Stop()
// Drive the timer from this goroutine only. backoff.ExponentialBackOff is
// not safe for concurrent use, so we never hand it to backoff.NewTicker's
// internal goroutine; instead we own every NextBackOff/Reset call here.
timer := time.NewTimer(backoffConfig.NextBackOff())
defer timer.Stop()

for {
select {
case <-g.stopCh:
g.logger.Debug("Worker stopping", zap.Int("worker_id", workerID))
return
case <-backoffTicker.C:
case <-timer.C:
if g.tracker != nil && !g.tracker.Acquire() {
select {
case <-g.stopCh:
return
case <-g.tracker.ResumeC():
timer.Reset(backoffConfig.NextBackOff())
continue
}
}
Expand All @@ -153,9 +157,11 @@ func (g *ApacheCombinedLogGenerator) worker(workerID int) {
g.logger.Error("Failed to write log",
zap.Int("worker_id", workerID),
zap.Error(err))
timer.Reset(backoffConfig.NextBackOff())
continue
}
backoffConfig.Reset()
timer.Reset(backoffConfig.NextBackOff())
}
}
}
Expand Down
12 changes: 9 additions & 3 deletions generator/apache_error/apache_error.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -128,20 +128,24 @@ func (g *ApacheErrorLogGenerator) worker(workerID int) {
backoffConfig.MaxInterval = 5 * time.Second
backoffConfig.MaxElapsedTime = 0 // Never stop retrying

backoffTicker := backoff.NewTicker(backoffConfig)
defer backoffTicker.Stop()
// Drive the timer from this goroutine only. backoff.ExponentialBackOff is
// not safe for concurrent use, so we never hand it to backoff.NewTicker's
// internal goroutine; instead we own every NextBackOff/Reset call here.
timer := time.NewTimer(backoffConfig.NextBackOff())
defer timer.Stop()

for {
select {
case <-g.stopCh:
g.logger.Debug("Worker stopping", zap.Int("worker_id", workerID))
return
case <-backoffTicker.C:
case <-timer.C:
if g.tracker != nil && !g.tracker.Acquire() {
select {
case <-g.stopCh:
return
case <-g.tracker.ResumeC():
timer.Reset(backoffConfig.NextBackOff())
continue
}
}
Expand All @@ -150,9 +154,11 @@ func (g *ApacheErrorLogGenerator) worker(workerID int) {
g.logger.Error("Failed to write log",
zap.Int("worker_id", workerID),
zap.Error(err))
timer.Reset(backoffConfig.NextBackOff())
continue
}
backoffConfig.Reset()
timer.Reset(backoffConfig.NextBackOff())
}
}
}
Expand Down
12 changes: 9 additions & 3 deletions generator/filegen/filegen.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -443,20 +443,24 @@ func (g *FileLogGenerator) worker(id int, files []string) {
backoffConfig.MaxInterval = 5 * time.Second
backoffConfig.MaxElapsedTime = 0 // Never stop retrying

backoffTicker := backoff.NewTicker(backoffConfig)
defer backoffTicker.Stop()
// Drive the timer from this goroutine only. backoff.ExponentialBackOff is
// not safe for concurrent use, so we never hand it to backoff.NewTicker's
// internal goroutine; instead we own every NextBackOff/Reset call here.
timer := time.NewTimer(backoffConfig.NextBackOff())
defer timer.Stop()

for {
select {
case <-g.stopCh:
g.logger.Debug("Worker received stop signal", zap.Int("id", id))
return
case <-backoffTicker.C:
case <-timer.C:
if g.tracker != nil && !g.tracker.Acquire() {
select {
case <-g.stopCh:
return
case <-g.tracker.ResumeC():
timer.Reset(backoffConfig.NextBackOff())
continue
}
}
Expand All @@ -471,11 +475,13 @@ func (g *FileLogGenerator) worker(id int, files []string) {
g.logger.Error("Error reading file", zap.String("file", file), zap.Error(err))
generator.BlitzGeneratorWriteErrorsCounter.Add(context.Background(), 1, componentName)
// On error, backoff will automatically handle retry timing
timer.Reset(backoffConfig.NextBackOff())
continue
}

// On success, reset backoff to configured rate
backoffConfig.Reset()
timer.Reset(backoffConfig.NextBackOff())
fileIdx += g.workers
}
}
Expand Down
12 changes: 9 additions & 3 deletions generator/json/json.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -183,20 +183,24 @@ func (g *JSONLogGenerator) worker(workerID int) {
backoffConfig.MaxInterval = 5 * time.Second
backoffConfig.MaxElapsedTime = 0 // Never stop retrying

backoffTicker := backoff.NewTicker(backoffConfig)
defer backoffTicker.Stop()
// Drive the timer from this goroutine only. backoff.ExponentialBackOff is
// not safe for concurrent use, so we never hand it to backoff.NewTicker's
// internal goroutine; instead we own every NextBackOff/Reset call here.
timer := time.NewTimer(backoffConfig.NextBackOff())
defer timer.Stop()

for {
select {
case <-g.stopCh:
g.logger.Debug("Worker stopping", zap.Int("worker_id", workerID))
return
case <-backoffTicker.C:
case <-timer.C:
if g.tracker != nil && !g.tracker.Acquire() {
select {
case <-g.stopCh:
return
case <-g.tracker.ResumeC():
timer.Reset(backoffConfig.NextBackOff())
continue
}
}
Expand All @@ -205,9 +209,11 @@ func (g *JSONLogGenerator) worker(workerID int) {
g.logger.Error("Failed to write log",
zap.Int("worker_id", workerID),
zap.Error(err))
timer.Reset(backoffConfig.NextBackOff())
continue
}
backoffConfig.Reset()
timer.Reset(backoffConfig.NextBackOff())
}
}
}
Expand Down
12 changes: 9 additions & 3 deletions generator/kubernetes/kubernetes.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -152,20 +152,24 @@ func (g *Generator) worker(workerID int) {
backoffConfig.MaxInterval = 5 * time.Second
backoffConfig.MaxElapsedTime = 0

backoffTicker := backoff.NewTicker(backoffConfig)
defer backoffTicker.Stop()
// Drive the timer from this goroutine only. backoff.ExponentialBackOff is
// not safe for concurrent use, so we never hand it to backoff.NewTicker's
// internal goroutine; instead we own every NextBackOff/Reset call here.
timer := time.NewTimer(backoffConfig.NextBackOff())
defer timer.Stop()

for {
select {
case <-g.stopCh:
g.logger.Debug("Worker stopping", zap.Int("worker_id", workerID))
return
case <-backoffTicker.C:
case <-timer.C:
if g.tracker != nil && !g.tracker.Acquire() {
select {
case <-g.stopCh:
return
case <-g.tracker.ResumeC():
timer.Reset(backoffConfig.NextBackOff())
continue
}
}
Expand All @@ -174,9 +178,11 @@ func (g *Generator) worker(workerID int) {
g.logger.Error("Failed to write log",
zap.Int("worker_id", workerID),
zap.Error(err))
timer.Reset(backoffConfig.NextBackOff())
continue
}
backoffConfig.Reset()
timer.Reset(backoffConfig.NextBackOff())
}
}
}
Expand Down
12 changes: 9 additions & 3 deletions generator/nginx/nginx.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -155,20 +155,24 @@ func (g *Generator) worker(workerID int) {
backoffConfig.MaxInterval = 5 * time.Second
backoffConfig.MaxElapsedTime = 0

backoffTicker := backoff.NewTicker(backoffConfig)
defer backoffTicker.Stop()
// Drive the timer from this goroutine only. backoff.ExponentialBackOff is
// not safe for concurrent use, so we never hand it to backoff.NewTicker's
// internal goroutine; instead we own every NextBackOff/Reset call here.
timer := time.NewTimer(backoffConfig.NextBackOff())
defer timer.Stop()

for {
select {
case <-g.stopCh:
g.logger.Debug("Worker stopping", zap.Int("worker_id", workerID))
return
case <-backoffTicker.C:
case <-timer.C:
if g.tracker != nil && !g.tracker.Acquire() {
select {
case <-g.stopCh:
return
case <-g.tracker.ResumeC():
timer.Reset(backoffConfig.NextBackOff())
continue
}
}
Expand All @@ -177,9 +181,11 @@ func (g *Generator) worker(workerID int) {
g.logger.Error("Failed to write log",
zap.Int("worker_id", workerID),
zap.Error(err))
timer.Reset(backoffConfig.NextBackOff())
continue
}
backoffConfig.Reset()
timer.Reset(backoffConfig.NextBackOff())
}
}
}
Expand Down
12 changes: 9 additions & 3 deletions generator/okta/okta.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -269,20 +269,24 @@ func (g *Generator) worker(workerID int) {
backoffConfig.MaxInterval = 5 * time.Second
backoffConfig.MaxElapsedTime = 0

backoffTicker := backoff.NewTicker(backoffConfig)
defer backoffTicker.Stop()
// Drive the timer from this goroutine only. backoff.ExponentialBackOff is
// not safe for concurrent use, so we never hand it to backoff.NewTicker's
// internal goroutine; instead we own every NextBackOff/Reset call here.
timer := time.NewTimer(backoffConfig.NextBackOff())
defer timer.Stop()

for {
select {
case <-g.stopCh:
g.logger.Debug("Worker stopping", zap.Int("worker_id", workerID))
return
case <-backoffTicker.C:
case <-timer.C:
if g.tracker != nil && !g.tracker.Acquire() {
select {
case <-g.stopCh:
return
case <-g.tracker.ResumeC():
timer.Reset(backoffConfig.NextBackOff())
continue
}
}
Expand All @@ -291,9 +295,11 @@ func (g *Generator) worker(workerID int) {
g.logger.Error("Failed to write log",
zap.Int("worker_id", workerID),
zap.Error(err))
timer.Reset(backoffConfig.NextBackOff())
continue
}
backoffConfig.Reset()
timer.Reset(backoffConfig.NextBackOff())
}
}
}
Expand Down
12 changes: 9 additions & 3 deletions generator/paloalto/paloalto.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -118,28 +118,34 @@ func (g *Generator) worker(workerID int) {
backoffConfig.MaxInterval = 5 * time.Second
backoffConfig.MaxElapsedTime = 0

backoffTicker := backoff.NewTicker(backoffConfig)
defer backoffTicker.Stop()
// Drive the timer from this goroutine only. backoff.ExponentialBackOff is
// not safe for concurrent use, so we never hand it to backoff.NewTicker's
// internal goroutine; instead we own every NextBackOff/Reset call here.
timer := time.NewTimer(backoffConfig.NextBackOff())
defer timer.Stop()

for {
select {
case <-g.stopCh:
g.logger.Debug("Worker stopping", zap.Int("worker_id", workerID))
return
case <-backoffTicker.C:
case <-timer.C:
if g.tracker != nil && !g.tracker.Acquire() {
select {
case <-g.stopCh:
return
case <-g.tracker.ResumeC():
timer.Reset(backoffConfig.NextBackOff())
continue
}
}
if err := g.generateAndWrite(workerID); err != nil {
g.logger.Error("Failed to write log", zap.Int("worker_id", workerID), zap.Error(err))
timer.Reset(backoffConfig.NextBackOff())
continue
}
backoffConfig.Reset()
timer.Reset(backoffConfig.NextBackOff())
}
}
}
Expand Down
12 changes: 9 additions & 3 deletions generator/postgres/postgres.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -307,20 +307,24 @@ func (g *Generator) worker(workerID int) {
backoffConfig.MaxInterval = 5 * time.Second
backoffConfig.MaxElapsedTime = 0

backoffTicker := backoff.NewTicker(backoffConfig)
defer backoffTicker.Stop()
// Drive the timer from this goroutine only. backoff.ExponentialBackOff is
// not safe for concurrent use, so we never hand it to backoff.NewTicker's
// internal goroutine; instead we own every NextBackOff/Reset call here.
timer := time.NewTimer(backoffConfig.NextBackOff())
defer timer.Stop()

for {
select {
case <-g.stopCh:
g.logger.Debug("Worker stopping", zap.Int("worker_id", workerID))
return
case <-backoffTicker.C:
case <-timer.C:
if g.tracker != nil && !g.tracker.Acquire() {
select {
case <-g.stopCh:
return
case <-g.tracker.ResumeC():
timer.Reset(backoffConfig.NextBackOff())
continue
}
}
Expand All @@ -329,9 +333,11 @@ func (g *Generator) worker(workerID int) {
g.logger.Error("Failed to write log",
zap.Int("worker_id", workerID),
zap.Error(err))
timer.Reset(backoffConfig.NextBackOff())
continue
}
backoffConfig.Reset()
timer.Reset(backoffConfig.NextBackOff())
}
}
}
Expand Down
Loading