[LU-9493] Recovered clients + evicted clients != total clients

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Affects Version/s: Lustre 2.8.0
• Severity: 3
• Assignee: Mikhail Pershin
• Reporter: Ned Bass

Description

After an MDT completed recovery the accounting of recovered clients appears to be off. Only 2651 of 2827 were recovered yet 0 were evicted. Is there a third category not listed that would account for the other 176 clients? Or is this a bookkeeping bug? Perhaps not coincidentally, we see 176 clients get evicted over two hours after recovery completed.

May 10 09:17:07 zinc1 kernel: Lustre: lsh-MDT0000: Will be in recovery for at least 5:00, or until 2827 clients reconnect
...
May 10 09:22:07 zinc1 kernel: Lustre: lsh-MDT0000: Recovery over after 5:01, of 2827 clients 2651 recovered and 0 were evicted.
...
May 10 11:30:36 zinc1 kernel: Lustre: lsh-MDT0000: haven't heard from client 81e4b56c-c961-07f6-b732-42fce10b4acf (at 192.168.120.150@o2ib20) in 227 seconds. I think it's dead, and I am evicting it. exp ffff887f33220c00, cur 1494441035 expire 1494440885 last 1494440808
May 10 11:30:36 zinc1 kernel: Lustre: Skipped 175 previous similar messages