For now, LFSCK scan reports status and information of the last run via sysfs (oi_scrub, lfsck_layout and lfsck_namespace) or "lfs lfsck_query". These report:

• counters like the number of entries scanned, failed or repaired
• LFSCK instance information: last time scan, LSFCK arguments etc ...
• status: incomplete, failed success etc...

But it missed detailed information about what has been fixed or not. Those information can be retrieved via debug log on each node:

$ clush -w@lustre_server "lctl set_param debug=+lfsck debug_mb=1024; lctl clear"
$ lctl lfsck_start -A
$ lctl lfsck_query -w
$ clush -w@lustre_server "lctl dk > /var/log/lfsck_report"

LFSCK debug logs are not automatically store and those are hard to read (not generic). For dry-run scan those logs are sometimes missing.

So it would be great to have a clear LFSCK reports to get what have been done or will be done (dry-run).