[autotest] Add a ping check to avoid ssh timeout
Collecting logs from dut requires ssh connection. When the dut is offline, the
ssh connection takes 1 minute to time out. crash collection makes multiple
log collections, and the timeouts add unnecessary overhead.
This change adds a cached ping status in ssh host object. The cache expires in
5 minutes by default. Log collection can be skipped if the cached status is
False.
BUG=None
TEST=local run repair
Change-Id: I23f84172f4ab3ea8a0e63c53afaf9ba139eb8455
Reviewed-on: https://chromium-review.googlesource.com/590484
Commit-Ready: Dan Shi <[email protected]>
Tested-by: Dan Shi <[email protected]>
Reviewed-by: Paul Hobbs <[email protected]>
diff --git a/server/site_crashcollect.py b/server/site_crashcollect.py
index 8a1bceb..b4ac6a0 100644
--- a/server/site_crashcollect.py
+++ b/server/site_crashcollect.py
@@ -202,6 +202,12 @@
if not os.path.exists(infodir):
os.mkdir(infodir)
orphans = []
+
+ if not host.check_cached_up_status():
+ logging.warning('Host %s did not answer to ping, skip fetching '
+ 'orphaned crashdumps.', host.hostname)
+ return orphans
+
try:
for file in _find_orphaned_crashdumps(host):
logging.info('Collecting %s...', file)
