FreeNAS

Also tested 11.2 and cannot reproduce. Attached are debugs from the systems I attempted.

I wrote the following script which did produce a hang in listing operations while the data set, and snapshot creation tasks where underway.

#!/bin/sh

if [ -z $1 ] ; then
  echo "Please provide pool name!" 
  exit
fi

pool=$1

# Create 20 datasets
for i in `seq 1 1000`; do
   zfs create ${pool}/$i
done

# Create 1000 snapshots
for i in `seq 1 10000`; do
    zfs snapshot ${pool}@$i
done

I did crash middleware once on an encrypted volume while trying multiple operations to ddos the machine at the same time. However with single operations to create datasets, and snapshots in batch the UI became responsive again.

I also created a dataset with an extremely long name like this:

`tank/reallylargedsnameisverylongonpurpose/CameCaseDSNAMEwith12345678/makingnameevenlongeronpurpose/letmakethenameevenlonger/howaboutmakingthenamethislong@auto-20190321.1321-2w`

I was of course given warning about character limit so i shortened to the limits, and I was still able to create the perodic snap task successfully using both django, and angular UI.

I found the issue on a call with Josh today. Let me know if you need a system setup for reproduction. Attached is an image showing autorepl.py trying to run and giving a traceback when a periodic snapshot is created. This is caused when a destination replication server goes offline. As a result of a replication server going offline no new periodic snapshots cannot be created. That is the bug we see here.

To reproduce:
Add a periodic snapshot
Add a replication task to a remote server
Take the remote server offline
Try to add another periodic snapshot task

11.3 PR: https://github.com/freenas/freenas/pull/2816