File tree Expand file tree Collapse file tree
Expand file tree Collapse file tree Original file line number Diff line number Diff line change @@ -247,12 +247,14 @@ def RemoteCommand(
247247
248248 Returns:
249249 A tuple of stdout and stderr from running the command.
250-
251- Raises:
252- RemoteCommandError: If there was a problem issuing the command.
253250 """
251+ slurm_path = (
252+ 'export PATH=$PATH:/usr/local/bin:/usr/local/sbin:'
253+ '/apps/slurm/current/bin:/apps/slurm/bin:'
254+ '/opt/slurm/bin:/usr/local/slurm/bin; '
255+ )
254256 return self .headnode_vm .RemoteCommand (
255- f'{ env } srun -N { self .num_workers } { command } ' ,
257+ f'{ slurm_path } { env } srun -N { self .num_workers } { command } ' ,
256258 ignore_failure = ignore_failure ,
257259 timeout = timeout ,
258260 ** kwargs ,
@@ -372,14 +374,19 @@ def Delete(self):
372374 @vm_util .Retry (
373375 fuzz = 0 ,
374376 timeout = 1800 ,
375- max_retries = 5 ,
377+ max_retries = 30 ,
376378 retryable_exceptions = (errors .Resource .RetryableCreationError ,),
377379 )
378380 def _WaitForClusterReady (self ):
379381 if self .unmanaged :
380382 return
383+ slurm_path = (
384+ 'export PATH=$PATH:/usr/local/bin:/usr/local/sbin:'
385+ '/apps/slurm/current/bin:/apps/slurm/bin:'
386+ '/opt/slurm/bin:/usr/local/slurm/bin; '
387+ )
381388 if not self .headnode_vm .TryRemoteCommand (
382- f'srun -N { self .num_workers } hostname'
389+ f'{ slurm_path } srun -N { self .num_workers } hostname'
383390 ):
384391 raise errors .Resource .RetryableCreationError ('Cluster not ready.' )
385392
Original file line number Diff line number Diff line change 77 region: {{ region }}
88 zone: {{ zone }}
99
10+ {% if 'staging' in zone %}
11+ # Add skip validators for staging
12+ validators:
13+ - {validator: test_zone_exists, skip: true}
14+ - {validator: test_zone_in_region, skip: true}
15+ - {validator: test_region_exists, skip: true}
16+ - {validator: test_machine_type_in_zone, skip: true}
17+ - {validator: test_disk_type_in_zone, skip: true}
18+
19+ # Add provider overrides for staging
20+ terraform_providers:
21+ google:
22+ source: hashicorp/google
23+ version: ">= 6.22.0"
24+ configuration:
25+ project: $(vars.project_id)
26+ region: $(vars.region)
27+ zone: $(vars.zone)
28+ compute_custom_endpoint: "https://compute.mtls.googleapis.com/compute/staging_v1/"
29+ google-beta:
30+ source: hashicorp/google-beta
31+ version: ">= 6.22.0"
32+ configuration:
33+ project: $(vars.project_id)
34+ region: $(vars.region)
35+ zone: $(vars.zone)
36+ compute_custom_endpoint: "https://compute.mtls.googleapis.com/compute/staging_v1/"
37+ {% endif %}
38+
1039deployment_groups:
1140- group: primary
1241 modules:
@@ -32,7 +61,10 @@ deployment_groups:
3261 enable_public_ips: true
3362 enable_oslogin: false
3463 machine_type: {{ worker_machine_type }}
35- disk_type: pd-balanced
64+ disk_type: {{ 'hyperdisk-balanced' if 'h4d' in worker_machine_type else 'pd-balanced' }}
65+ {% if 'h4d' in worker_machine_type %}
66+ on_host_maintenance: TERMINATE
67+ {% endif %}
3668 bandwidth_tier: gvnic_enabled
3769 allow_automatic_updates: false
3870 instance_image_custom: true
You can’t perform that action at this time.
0 commit comments