fix(dgraph): ludicrous mode mutation error

[harshil-goel]

https://discuss.dgraph.io/t/ludicrous-mode-fails-queries-on-21-million-dataset/7320/2

Time taken to live load 21 million dataset in ludicrous mode went down from 5:11 min on average to 3:01 min on average.

---

This change is 

Docs Preview:

[jarifibrahim]

I don't understand everything but based on https://discuss.dgraph.io/t/ludicrous-mode-fails-queries-on-21-million-dataset/7320 and the explanation @harshil-goel gave me, this change looks good to me.

Reviewed 1 of 5 files at r2.
Reviewable status: 1 of 5 files reviewed, 4 unresolved discussions (waiting on @harshil-goel, @manishrjain, and @vvbalaji-dgraph)

---

systest/21million/test-21million.sh, line 61 at r2 (raw file):

    --mode          normal = run dgraph in normal mode
                    none = run dgraph in normal mode
		    ludicrous = run dgraph in ludicrous mode

Indentation.

---

worker/draft.go, line 646 at r2 (raw file):

			totalSize += int64(psz)

			if x.WorkerConfig.LudicrousMode && proposal.Mutations != nil && proposal.Mutations.StartTs == 0 {

If we're not assigning startTs to mutations, can they still have a non-zero startTs? If not, I'd convert this proposal.Mutations.StartTs == 0 into an assert inside the if

---

worker/draft.go, line 1164 at r2 (raw file):

							span.Annotate(nil, "Proposal found in CommittedEntries")
						}
						if x.WorkerConfig.LudicrousMode && len(proposal.Mutations.GetEdges()) > 0 {

I believe this change is not related to this PR.

---

worker/task.go, line 2185 at r2 (raw file):

	for itr.Seek(countKey); itr.Valid(); itr.Next() {
		item := itr.Item()
		key := append(make([]byte, 0), item.Key()...)

I believe this change is not related to this PR.

[harshil-goel]

Reviewable status: 1 of 5 files reviewed, 4 unresolved discussions (waiting on @jarifibrahim, @manishrjain, and @vvbalaji-dgraph)

---

systest/21million/test-21million.sh, line 61 at r2 (raw file):

Previously, jarifibrahim (Ibrahim Jarif) wrote…

Indentation.

Done.

---

worker/draft.go, line 646 at r2 (raw file):

Previously, jarifibrahim (Ibrahim Jarif) wrote…

If we're not assigning startTs to mutations, can they still have a non-zero startTs? If not, I'd convert this proposal.Mutations.StartTs == 0 into an assert inside the if

ALTER Operations already have a timestamp. do end up coming with startTs. The same check == 0 is applied when we assign a startTs to a mutation.

---

worker/draft.go, line 1164 at r2 (raw file):

Previously, jarifibrahim (Ibrahim Jarif) wrote…

I believe this change is not related to this PR.

I needed the change to run the tests. I will get merge those other PRs first.

---

worker/task.go, line 2185 at r2 (raw file):

Previously, jarifibrahim (Ibrahim Jarif) wrote…

I believe this change is not related to this PR.

I needed the change to run the tests. I will get merge those other PRs first.

[manishrjain]

Reviewed 4 of 5 files at r2, 1 of 1 files at r3.
Reviewable status: all files reviewed, 4 unresolved discussions (waiting on @jarifibrahim and @vvbalaji-dgraph)

---

worker/draft.go, line 646 at r2 (raw file):

Previously, harshil-goel (Harshil Goel) wrote…

ALTER Operations already have a timestamp. do end up coming with startTs. The same check == 0 is applied when we assign a startTs to a mutation.

100 chars. Seems like a fine thing to do.