fix tar_extract task: switch LOCAL to HDD disk, bump cpu and preemptible

The tar_extract task was failing 100% on GCP Batch with error "The job
was stopped before the command finished. GCP Batch task exited with
Success(0)." The LOCAL disk type requires NVMe local SSDs which are
incompatible with the small machine types GCP Batch selects for a
1-CPU/2-GB task. The task never started executing (no stderr/stdout
produced). Switching to HDD persistent disk resolves the provisioning
incompatibility. Also bump cpu to 2 and preemptible to 2 for better
scheduling compatibility and resilience.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: dpark01

pipes/WDL/tasks/tasks_utils.wdl

@@ -352,11 +352,11 @@ task tar_extract {
     runtime {
         docker: "quay.io/broadinstitute/viral-ngs:3.0.10-baseimage"
         memory: "2 GB"
-        cpu:    1
-        disks: "local-disk ~{disk_size} LOCAL"
+        cpu:    2
+        disks: "local-disk ~{disk_size} HDD"
         disk: "~{disk_size} GB" # TES
         dx_instance_type: "mem1_ssd1_v2_x2"
-        preemptible: 1
+        preemptible: 2
     }
     output {
         Array[File] files = glob("unpack/*")