Raft buffer is full error

[anahit-martirosyan]

When performing INSERT or DELETE operations with several vertices/edges I keep getting 'raft buffer is full' error, although there are no other reads or writes to the space.

I have setup a Nebula graph cluster with 3 nodes (3 metad services, 3 graphd services, 3 storaged services). I have create a space with the following configurations:
CREATE SPACE test_space (partition_num = 120, replica_factor = 3, vid_type = FIXED_STRING(100));

I have filled the space with some data. Then I needed to remove all the vertices of type 'x' with their corresponding edges. For this I performed the following command:
LOOKUP ON x YIELD id(vertex) as id | DELETE VERTEX $-.id WITH EDGE
As a result to this operation I got Graph memory exceeded error:

[ERROR (-1005)]: GraphMemoryExceeded: (-2600)

After this when I try to perform the same operation but with LIMIT, (or other INSERT or DELETE operations with several vertices/edges):
LOOKUP ON device_list YIELD id(vertex) as id | LIMIT 10 | DELETE VERTEX $-.id WITH EDGE
I get raft buffer is full error:

[ERROR (-1005)]: Storage Error: Part 10 raft buffer is full. Please retry later.

Can someone please help to understand what does 'raft buffer is full' error mean? Is there a way to flush raft buffer?