feat(multi-tenancy): make drop data namespace aware

[NamanJain8]

This PR makes DropData namespace-aware in a multi-tenant system. Now, guardians of a namespace can call drop data on their namespace. Earlier only the guardian of the galaxy was allowed to do cluster-wide drop operation.

---

This change is 

[manishrjain]

Reviewed 10 of 10 files at r1.
Reviewable status: all files reviewed, 4 unresolved discussions (waiting on @NamanJain8 and @vvbalaji-dgraph)

---

edgraph/access.go, line 49 at r1 (raw file):

}

func upsertGuardianAndGroot(closer *z.Closer, ns uint64) {

// do nothing.

---

worker/cdc_ee.go, line 123 at r1 (raw file):

	for ts, events := range cdc.pendingTxnEvents {
		if len(events) > 0 && binary.BigEndian.Uint64(events[0].Meta.Namespace) == ns {
			delete(cdc.pendingTxnEvents, ts)

Don't need to delete any events before drop all.

---

worker/cdc_ee.go, line 266 at r1 (raw file):

						return
					}
					cdc.resetPendingEventsForNs(ns)

no need.

---

worker/draft.go, line 329 at r1 (raw file):

		// Clear entire cache.
		// TODO(Naman): Would it be okay for now to reset cluster wide cache on drop data.

We don't want to drop entire clusters cache, just due to one namespace.

[NamanJain8]

Dismissed @aman-bansal and @manishrjain from 3 discussions.
Reviewable status: 7 of 10 files reviewed, 2 unresolved discussions (waiting on @aman-bansal, @manishrjain, and @vvbalaji-dgraph)

---

edgraph/access.go, line 49 at r1 (raw file):

Previously, manishrjain (Manish R Jain) wrote…

// do nothing.

Done.

---

worker/cdc_ee.go, line 123 at r1 (raw file):

Previously, manishrjain (Manish R Jain) wrote…

Don't need to delete any events before drop all.

As discussed with @aman-bansal , these events are cleared on the commit of the transaction. Hence, we need to clear out all these events, else they would stay in memory.

---

worker/cdc_ee.go, line 266 at r1 (raw file):

Previously, manishrjain (Manish R Jain) wrote…

no need.

Replied above.

---

worker/cdc_ee.go, line 433 at r1 (raw file):

Previously, aman-bansal (aman bansal) wrote…

please don't panic

Done.

---

worker/draft.go, line 329 at r1 (raw file):

Previously, manishrjain (Manish R Jain) wrote…

We don't want to drop entire clusters cache, just due to one namespace.

Done.