Marking 100 tasks as ‘failed’ takes about 30-40 seconds. Could use some optimization in your database commit.
Hmm, that’s fairly worrisome…
Hey guys,
The task operations definitely need some optimization. We wanted to get those features back in and optimize them during the 6.0 beta. It’s logged as a bug, and we will definitely be taking a look at this.
Cheers,
- Ryan
Hmm, I checked this code, and it is pretty well optimized. I also can’t seem to reproduce it here.
You’ve moved your database to 64-bit right? Do you still have this performance issue since moving to 64-bit?
Nope still 32 bit.
Oh man. Please (!!) move to a 64 bit Mongo install if possible. 
We essentially will not be recommending or supporting 32 bit mongo servers in production environment going forward. We’ll still be shipping the 32 bit mongo install so that it can be used for testing, but the installer will show a big warning now if you try to install 32 bit mongo.
If/when you do get a chance to move to 64 bit, try marking 100+ tasks as failed again and let us know if it’s still slow.
If any other beta testers else can test this against 64 bit mongo, we would love to hear about the performance you experienced.
Cheers,
- Ryan
I just requeued a job. Then marked its tasks as failed. Instantaneous.
Could it be locking up if a task is maybe currently rendering for that task to stop before returning the UI?
EDIT: to answer my question: no. I failed a rendering task along with a dozen others and it was still instant.