Both cbbackup and webgui hangs

I’m trying to use cbbackup to backup an entire cluster (9 buckets, less than 1 GB total size), but when it comes to one of the buckets the program just freezes, no more output is being generated, even with -vvv.

The bucket that’s causing this is just 664 docs, other buckets with over 500k docs works without a hitch. I noticed as well that if I use the webgui to view the bucket, when I get to a certain point, the chrome tab crashes, about 75% in (which is ~ how far the backup goes). However it works in Firefox (with a short freeze when loading the page).

The only thing I have been able to figure out is that I think the documents causing this is larger than most other docs in the bucket, From the network monitoring in FF the other pages are around 30-77 KB while the ones causing issues are around 1 MB.

I’m using Version: 3.0.1 Enterprise Edition (build-1444)

I first though that the problem was particularly large documents, but after generating some megabytes of json and inserting into a test bucket and backing it up I abandoned this though. I then tried updating the server from 3.0.1 (1444) to 3.0.2(1630), but this didn’t change anything either.

The only thing I have gotten working is to create a new identical bucket, clone all docs from the original bucket to the new one and then backing that up. So it must be something wonky with that one bucket. Although when I do this cbbackup starts working, but chrome still crashes.

So I’m probably leaving this for now, but please post any findings anyone else might have.

Ok this is seriously a severe issue. I have seen this happening to more and more buckets, seemingly random. Some with ~25k documents and others with 3 docs. This is now preventing us from performing any kind of backup.
As per my previous message this is not something document or view specific since it works if I just clone a bucket to a new one.

Edit: I am also seeing that this issue completely prevents rebalancing. It goes to a certain percentage and then freezes.

I would probably recommend posting a ‘collect info’ which can be done from the Web UI and then filing an issue. Since it’s an intermittent problem, something in the logs will probably make it clear.

Of course, if you have an Enterprise Edition subscription, you can also contact support to review the case and track it to resolution (there’s no guaranteed service level with filing an issue).