Live Loader aborts/retries take a long time and a lot of space

Issues Dgraph
1

@minhaj

hi, may I ask what’s the “arising conflicts” ?

now I also encountered this situation with live：

[23:18:37Z] Elapsed: 15h11m55s Txns: 2402963 N-Quads: 2402947802 N-Quads/s [last 5s]:     0 Aborts: 338
[23:18:42Z] Elapsed: 15h12m00s Txns: 2402963 N-Quads: 2402947802 N-Quads/s [last 5s]:     0 Aborts: 338
[23:18:47Z] Elapsed: 15h12m05s Txns: 2402963 N-Quads: 2402947802 N-Quads/s [last 5s]:     0 Aborts: 338
[23:18:52Z] Elapsed: 15h12m10s Txns: 2402963 N-Quads: 2402947802 N-Quads/s [last 5s]:     0 Aborts: 338

The same debug has been going on for an hour.
I checked the hard disk space and found that the utilization rate was 100%. So I deleted some files. Now the utilization rate is 96% (this also has about 30g of space).

However, the same debug has been used in the past half an hour.
So isn’t the problem due to insufficient space? What caused the conflict?

2

I believe this is something unrelated. I never got a last 5s: 0 report.

I think after I found my conflicts were due to bad rdf statements from my script trying to link data using blank nodes that were contained a typo. Not sure exactly, but that is my guess.

3

I’ve moved this to a new issue. I think what @youyin123 is asking about is whether it seems right that the abort/retry takes 5 seconds and uses a lot of disk space;

1 Like
4

@chewxy

I’m not sure it was because the abort/retry uses a lot of disk space, because the load data, json file, size is 219G.
until I killed the live, the p folder size is 572G, Is it normal？