fix(fetch_all_rows): pass down timeout correctly

`fetch_all_rows` was passing timeout only to `request_all` call
while the inital request of the first page was using the default
timeout which isn't always enough, and can lead to the following
error:
```
(TestFrameworkEvent Severity.ERROR) period_type=one-time
event_id=e2d16764-de17-4150-bc90-ba2c6f5e5b90,
source=PartitionsValidationAttributes
message=Failed to collect partition info. Error details:
Requested pages were not delivered before timeout. Requested: 1; retrieved: 0; empty retrieved 0
```

in this change we change the session default timeout, assuming
this session is used mostly for those calls

sdcm/utils/database_query_utils.py

@@ -209,6 +209,7 @@ def fetch_all_rows(session, default_fetch_size, statement, retries: int = 4, tim
         LOGGER.debug("Fetch all rows by statement: %s", statement)
     session.default_fetch_size = default_fetch_size
     session.default_consistency_level = ConsistencyLevel.QUORUM
+    session.default_timeout = timeout
 
     @retrying(n=retries, sleep_time=5, message='Fetch all rows', raise_on_exceeded=raise_on_exceeded)
     def _fetch_rows() -> list: