Ceph

related thread

<sjust> I'm not sure I understand your question
<sjust> it shouldn't happen with default settings
<loicd> that's what confused me I guess. When can it happen ? 
<sjust> but if you turn the min log entries option to a small number, the primary might wake up after a brief nap to find that all of its entries are divergent
<sjust> although now it seems like that should cause backfill
<sjust> hmm
* loicd thinking
<sjust> ah, if log.tail == auth_log.head
<loicd> but
<loicd> assert(newhead > log.tail);
<loicd> would throw then, right ? 
<sjust> yeah, you are right
<sjust> I don't think that case can happen
* loicd digging the history of changes for a hint 
<sjust> loicd: I think there may be a conflict with calc_acting where it chooses which peers can be activated, you may want to make a bug
<loicd> ok, sjust I'll check this tomorrow morning. Thanks for the advice :-)
<sjust> I think the assert is actually wrong, but we'll never see the problem case due to osd_min_pg_log_entries
<loicd> the history goes straight back to when rewind_divergent_log was not yet a function on its own, difficult to compare
<sjust> yeah, the history probably won't be a lot of help
<loicd> not sure to understand why osd_min_pg_log_entries makes it impossible for newhead to be <= log.tail 
* loicd thinking
<sjust> not impossible, but very unlikely
* loicd contemplates
<loicd>  // The logs must overlap.
<loicd>   assert(log.head >= olog.tail && olog.head >= log.tail);
<sjust> right, the question is whether the constraint is
<sjust> log.head >= olog.tail
<sjust> or
<sjust> log.head > olog.tail
<sjust> in the latter case, there may be no entries in common
<sjust> in the former case, there must be at least 1 entry in common
<loicd> so the assert can only be triggered if olog.head == log.tail because of the previous assert on  olog.head >= log.tail)