tests: compare digests in test_peer_recovery (#10853)

## Problem

Test fails when comparing the first WAL segment because the system id in
the segment header is different. The system id is not consistently set
correctly since segments are usually inited on the safekeeper sync step
with sysid 0.

## Summary of Chnages

Compare timeline digests instead. This skips the header.

Closes https://github.com/neondatabase/neon/issues/10596

test_runner/regress/test_wal_acceptor.py

@@ -1445,6 +1445,7 @@ def test_peer_recovery(neon_env_builder: NeonEnvBuilder):
 
     # roughly fills one segment
     endpoint.safe_psql("insert into t select generate_series(1,250000), 'payload'")
+    lsn = Lsn(endpoint.safe_psql("SELECT pg_current_wal_flush_lsn()")[0][0])
 
     endpoint.stop()  # stop compute
 
@@ -1473,7 +1474,15 @@ def test_peer_recovery(neon_env_builder: NeonEnvBuilder):
         "flush_lsn to get aligned",
     )
 
-    cmp_sk_wal([sk1, sk2], tenant_id, timeline_id)
+    sk1_digest = sk1.http_client().timeline_digest(
+        tenant_id, timeline_id, sk1.get_timeline_start_lsn(tenant_id, timeline_id), lsn
+    )
+
+    sk2_digest = sk1.http_client().timeline_digest(
+        tenant_id, timeline_id, sk2.get_timeline_start_lsn(tenant_id, timeline_id), lsn
+    )
+
+    assert sk1_digest == sk2_digest
 
     # stop one of safekeepers which weren't recovering and insert a bit more to check we can commit
     env.safekeepers[2].stop()