"short write" errors

[bploetz]

We started seeing this all of a sudden last night out of the blue:

error enqueueing update-profile job: error enqueueing job: short write
…cessors.(*Processor).HandleError (/opt/reward-service/processors/processor.go:34)
…lReceiptProcessor).Process (/opt/reward-service/processors/digital_receipt.go:230)
….func1 (/go/pkg/mod/github.com/contribsys/faktory_worker_go@v1.6.0/manager.go:47)
…tch (/go/pkg/mod/github.com/contribsys/faktory_worker_go@v1.6.0/middleware.go:19)
…cessOne (/go/pkg/mod/github.com/contribsys/faktory_worker_go@v1.6.0/runner.go:139)
…process (/go/pkg/mod/github.com/contribsys/faktory_worker_go@v1.6.0/runner.go:95)
                         runtime.goexit (/usr/local/go/src/runtime/asm_amd64.s:1581)

The code in question:

func enqueueUpdateProfileJob(newRelicTxn *newrelic.Transaction, receiptID string, jobClient jobs.Client) error {
	defer newRelicTxn.StartSegment("enqueueUpdateProfileJob").End()
	jobArgument := make(map[string]interface{})
	jobArgument["id"] = receiptID
	jobArgs := []interface{}{jobArgument}
	retry := 5
	updateProfileJob := faktory.NewJob("update-profile", jobArgs...)
	updateProfileJob.Queue = "profiles"
	updateProfileJob.ReserveFor = 60
	updateProfileJob.Retry = &retry

	jpperr := jobClient.Push(updateProfileJob)
	if jpperr != nil {
		return fmt.Errorf("error enqueueing job: %v", jpperr)
	}

	return nil
}

[mperham]

Can you provide the backtrace for jpperr?

[mperham]

This can happen if two goroutines are sharing the same jobClient.

https://stackoverflow.com/questions/66174520/how-to-solve-short-write-error-when-writing-to-csv