Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
OpenDAS
dynamo
Commits
a0a09df0
Unverified
Commit
a0a09df0
authored
Apr 30, 2025
by
ishandhanani
Committed by
GitHub
Apr 30, 2025
Browse files
docs: add an example on how to use `--service-name` flag to spin up a standalone service (#915)
parent
2d39ded6
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
14 additions
and
1 deletion
+14
-1
examples/llm/multinode-examples.md
examples/llm/multinode-examples.md
+14
-1
No files found.
examples/llm/multinode-examples.md
View file @
a0a09df0
...
@@ -58,6 +58,19 @@ cd $DYNAMO_HOME/examples/llm
...
@@ -58,6 +58,19 @@ cd $DYNAMO_HOME/examples/llm
dynamo serve components.prefill_worker:PrefillWorker
-f
./configs/multinode-405b.yaml
dynamo serve components.prefill_worker:PrefillWorker
-f
./configs/multinode-405b.yaml
```
```
**Step 7**
: [Optional] Start more decode workers on other nodes
This example can be extended to more nodes as well. For example, if you'd like to spin up another decode worker, you can use
```
bash
# node X
export
NATS_SERVER
=
'<your-nats-server-address>'
# note this should start with nats://...
export
ETCD_ENDPOINTS
=
'<your-etcd-endpoints-address>'
cd
$DYNAMO_HOME
/examples/llm
dynamo serve components.worker:VllmWorker
-f
./configs/multinode-405b.yaml
--service-name
VllmWorker
```
Note the use of
`--service-name`
. This will only spin up the worker that you are requesting and ignore any
`depends`
statements.
### Client
### Client
In another terminal:
In another terminal:
...
@@ -213,4 +226,4 @@ dynamo serve components.prefill_worker:PrefillWorker -f ./configs/mutinode_disag
...
@@ -213,4 +226,4 @@ dynamo serve components.prefill_worker:PrefillWorker -f ./configs/mutinode_disag
In another terminal, you can send the same curl request as described in
In another terminal, you can send the same curl request as described in
[
aggregated deployment
](
#aggregated-deployment
)
, addressing to the ip of
[
aggregated deployment
](
#aggregated-deployment
)
, addressing to the ip of
the decode node.
the decode node.
\ No newline at end of file
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment