Query Result of the Variants of example too slow

Users
1

I modified the Tips and Tricks example query,

{
  test(func: has(director.film), first: 1) {
    uid
    expand(_all_) 
  }
}

query too slow,
Server latency: 1~16s
Rendering latency: 25ms,why?
I am not very clear here, please answer.

2

Hard to say, must be something in your context.

If you run this query against Dgraph’s play server. You gonna have:

Server latency: 56ms
Network latency: 240ms

Check you env, export your DB and upgrade to a latest Dgraph.

Cheers.

3

At present, I am using three HDD hard disk servers to build a cluster. The main Alpha is accessed through Ratel. However, the effect is still not improved. Is it that the query speed caused by the hard disk is several times different? this is too scary

4

HDD is a bit slow, but it is not so to the point of passing 1 second of latency. But it can be a problem depending on the physical state of the HDDs. They may delay some Dgraph’s background Tasks.

We recommend SSDs or NVME. Because the IOPs and access is much faster. A significant amount of Memory Cache also helps. You can use one Alpha with SSD (Also great amount of RAM) and the rest with HDDs tho.

Observe, in your case I do not know what to say, because I do not have access to its context. A lot things may be causing this.

5

THX，on the other hand,when importing data in bulk and live, I found that there are references to External IDs in the mutation.

But,as of version 0.8 Dgraph doesn’t natively support such external IDs as node identifiers.

But I am doing this by using bulk to import the discovery data or using the external ID. How does Dgraph map to the internal uid? Also, I use live and combine the Blank Nodes and UID methods and find that I will add attributes to my existing uid! Shouldn’t Blank Nodes create a blank node to put in the newly inserted data? PS: The data I started with was imported via bulk.

6

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.