prometheus/node_exporter · error

failed to retrieve bcache stats

Error message

failed to retrieve bcache stats: %w

What it means

bcacheCollector.Update calls c.fs.Stats() (or StatsWithoutPriority() when --collector.bcache.priority-stats is off) and wraps any read error in "failed to retrieve bcache stats". This means the sysfs bcache tree existed at construction but could not be read/walked during the scrape. The whole scrape for this collector fails and no bcache metrics are produced for that cycle.

Solutions

1. Re-check that the bcache device still exists under /sys/fs/bcache at scrape time (ls /sys/fs/bcache)
2. Check dmesg for bcache or block device errors and re-register the backing device if it was detached
3. Verify file permissions/readability for the node_exporter user on the sysfs bcache files
4. Update procfs dependency/kernel expectations if attributes were renamed in a newer kernel

Example fix

// before
 if err := c.Update(ch); err != nil {
	logger.Error(err.Error())
 }
// after
 if err := c.Update(ch); err != nil {
	if os.IsNotExist(errors.Unwrap(err)) {
		logger.Warn("bcache device removed; skipping scrape")
	} else {
		logger.Error(err.Error())
	}
 }

if fi, err := os.Stat("/sys/fs/bcache"); err != nil || len(dirEntries) == 0 { /* skip scrape */ }

err := collector.Update(ch)
if err != nil {
	var pe *os.PathError
	if errors.As(err, &pe) && errors.Is(pe.Err, syscall.ENODEV) {
		logger.Warn("bcache device detached; will retry next scrape")
	}
}

When it happens

Trigger: A periodic scrape where the /sys/fs/bcache directory disappeared (device detached), a backing device was hot-removed mid-walk, an underlying sysfs file read returned EIO/EACCES, or StatsWithoutPriority was selected and its files are unreadable.

Common situations: Hot-unplugging an SSD used as bcache cache during scraping; containers where sysfs files become unreadable; kernel version differences removing expected bcache attributes; I/O errors on the sysfs backing store.

Understand the failure class

Background: "failed to read file", EACCES, ENOENT and "could not read <path>" errors: when a program can't read a file from disk — this error's family across 49 libraries.

Related errors

• failed to open sysfs
• failed to retrieve bcachefs stats
• failed to retrieve Btrfs stats from procfs
• failed to scan DM-multipath devices
• couldn't get NUMA meminfo

AI-assisted analysis of prometheus/node_exporter@17ddd77c59 (2026-09-07). Data as JSON: /api/errors/94ccbbf4799c06ee. Report an issue: GitHub.

Appendix: source

Thrown at collector/bcache_linux.go:66

	return &bcacheCollector{
		fs:     fs,
		logger: logger,
	}, nil
}

// Update reads and exposes bcache stats.
// It implements the Collector interface.
func (c *bcacheCollector) Update(ch chan<- prometheus.Metric) error {
	var stats []*bcache.Stats
	var err error
	if *priorityStats {
		stats, err = c.fs.Stats()
	} else {
		stats, err = c.fs.StatsWithoutPriority()
	}
	if err != nil {
		return fmt.Errorf("failed to retrieve bcache stats: %w", err)
	}

	for _, s := range stats {
		c.updateBcacheStats(ch, s)
	}
	return nil
}

type bcacheMetric struct {
	name            string
	desc            string
	value           float64
	metricType      prometheus.ValueType
	extraLabel      []string
	extraLabelValue string
}

func bcachePeriodStatsToMetric(ps *bcache.PeriodStats, labelValue string) []bcacheMetric {

View on GitHub (pinned to 17ddd77c59)