Increase the initial inventory wait time of conductor

Some virtual subclouds failed to bootstrap when network
conditions, e.g. limited bandwidth, long delay, jitter and packet
loss, are applied. Under these conditions, inventory complete
notification from sysinv agent to sysinv conductor would arrive
a couple of seconds after the inventory timeout (90s) causing
subcloud bootstrap to fail.

This commit increases the conductor wait time by an additional
30 seconds to accommodate large batch subcloud deployment with
extended delay on mgmt network.

Test Plan:
  - Verify successful batch subcloud deployment with long delay
    on both oam and mgmt networks (up to 150ms).

Closes-Bug: 1998724
Change-Id: Iab65202eb6c5bed6969b002b5b1fb12729c371e6
Signed-off-by: Tee Ngo <tee.ngo@windriver.com>
This commit is contained in:
Tee Ngo 2022-12-04 19:06:48 -05:00
parent d7dede71ce
commit 9a97b8d83b
1 changed files with 1 additions and 1 deletions

View File

@ -1967,7 +1967,7 @@ ANSIBLE_BOOTSTRAP_FLAG = os.path.join(tsc.VOLATILE_PATH, ".ansible_bootstrap")
ANSIBLE_BOOTSTRAP_COMPLETED_FLAG = os.path.join(tsc.CONFIG_PATH,
".bootstrap_completed")
UNLOCK_READY_FLAG = os.path.join(tsc.PLATFORM_CONF_PATH, ".unlock_ready")
INVENTORY_WAIT_TIMEOUT_IN_SECS = 90
INVENTORY_WAIT_TIMEOUT_IN_SECS = 120
DEFAULT_RPCAPI_TIMEOUT_IN_SECS = 60
ANSIBLE_RESTORE_ROOK_FLAG = os.path.join(tsc.VOLATILE_PATH, ".ansible_restore_rook")