Items expired unexpected and some weirdness in stats

1. ./cluster_run -n 6
2. ./testrunner -i local.ini -t xdcr.uniXDCR.unidirectional.load_with_ops,items=50000,standard_buckets=1,sasl_buckets=1,ctopology=chain,rdirection=unidirection,replication_type=xmem,doc-ops=update-delete

two cluster - C1: n_0 and n_1
C2: n_2 and n_3
three buckets created on both C1 and C2 - default, standard_bucket_1, sasl_bucket_1
xdcr replications : C1 default -> C2 default
C1 standard_bucket_1 -> C2 standard_bucket_1
C1 sasl_bucket_1 -> C2 sasl_bucket_1
load 50000 items (with no expiration) on default, standard_bucket_1, sasl_bucket_1 on C1

After the test run, observe on C2 sasl_bucket_1 has only 39500 items.
In stats.log of n_2 and n_3, we see

curr_items: 17500
curr_items_tot: 39500

ep_num_ops_set_meta: 17500

vb_active_curr_items: 17500
vb_active_eject: 0
vb_active_expired: 7500

• Why there are 7500 documents expired?
• Why the value of ep_num_ops_set_meta is not equal to vb_active_curr_items + vb_active_expired?

testrunner log - https://friendpaste.com/5Vrrl1ZDyDIm1mOPhQi1ni