Chrome OS VM's /mnt/stateful_partition/ doesn't have enough space to deploy debug-built Chrome with --nostrip option.
This problem sometimes causes an error when Chrome tries to write something to disk.

Since this VM is also used in Chrome CQ for DCHECK-enabled Chrome, it can be a blocker for adding more ChromeOS tests.
Example:
crrev.com/c/1367333 is a CL to enable WebRTC tast tests on CQ bot, but new tests failed on Chrome infra with an error message like:
> database.cc(1808)] SyncDirectory sqlite error 13, errno 0: database or disk is full, sql: COMMIT
https://isolateserver.appspot.com/browse?namespace=default-gzip&digest=bade22ad814503bc09be35d48cb07689b9b73bac&as=chrome.PREVIOUS

The cause of this failure seems to be the lack of disk space under /mnt/stateful_partition/.
We could reproduce this disk-full error locally when by following  crbug.com/909975#c13  after cherry-picking crrev.com/c/1367388.
Without --nostrip for deploy_chrome, the test successfully worked locally.