Jeff Olivier

Hmm, if I add an ABT_thread_yield, in the thread creation loop, it is much better. So I suppose it's not a leak but may indicate a peak memory usage issue.

@carns For the example above, it makes no difference what I set that to. We can try in DAOS and see if it makes any difference. I'm trying to understand...

@carns in looking at the code, it looks like the max is 1024 now ``` #define ABTD_MEM_MAX_NUM_STACKS 1024 ... const uint32_t default_mem_max_stacks = ABTU_min_uint32(ABTD_MEM_MAX_TOTAL_STACK_SIZE / p_global->thread_stacksize, ABTD_MEM_MAX_NUM_STACKS); ``` 65536 could...

Also, it calls MPI_Abort which causes another error because MPI_Init hasn't been called yet.

The ticket isn't on the rc1 list

sorry, @mchaarawi I noticed I didn't like the if statement after removing the API call

> It looks to me that "obj discard" is only used in two cases today: > > 1. Discard old data on the target to be reintegrated before full reintegration....

> I'm wondering should we move further on this direction? > > Looks such fix won't solve the long rebuild reclaim time issue when the system has only few giant...

Note to @daos-stack/daos-gatekeeper this one is on top of #14382 so that should land first

sorry for the re-review. I tried an experiment and it failed. I fixed another issue in another PR and based this one on it. That preserved the reviews but after...