Merge #155535 #155544

155535: roachtest: deflake azure backup fixture roachtest r=dt a=msbutler

I suspect our azure external storage reader doesn't currently retry read timeouts. This patch disables the timeout to see if that deflakes the test. Once I've confirmed the root cause, i'll think through a fix.

Informs #155427

Release note: none

155544: kvclient/rangefeed: use larger machine for test runner r=stevendanna a=wenyihu6

This commit updates kv/kvclient/rangefeed to use a larger machine for its test
runner under race, as TestRangefeedCatchupStarvation previously hit an OOM under
race and appears to be related to an infra flake.

Fixes: #155334
Epic: none
Release note: none

Co-authored-by: Michael Butler <butler@cockroachlabs.com>
Co-authored-by: wenyihu6 <wenyi@cockroachlabs.com>

Author: 3 people

pkg/cmd/roachtest/tests/backup_fixtures.go

@@ -195,6 +195,7 @@ func (bd *backupDriver) prepareCluster(ctx context.Context) {
 				"kv.snapshot_rebalance.max_rate":                    "256 MiB",
 				"server.debug.default_vmodule":                      "s3_storage=2",
 				"cloudstorage.s3.client_retry_token_bucket.enabled": "false",
+				"cloudstorage.azure.try.timeout":                    "0s",
 			},
 			install.EnvOption{
 				fmt.Sprintf("COCKROACH_AZURE_APPLICATION_CREDENTIALS_FILE=%s", azureCredentialsFilePath),

pkg/kv/kvclient/rangefeed/BUILD.bazel

@@ -60,6 +60,10 @@ go_test(
         ":mock_rangefeed",  # keep
     ],
     embed = [":rangefeed"],
+    exec_properties = select({
+        "//build/toolchains:is_heavy": {"test.Pool": "large"},
+        "//conditions:default": {"test.Pool": "default"},
+    }),
     deps = [
         "//pkg/base",
         "//pkg/keys",