RDKEMW-20801: widen OCDM auto-start window to ~200s (-r 250 -d 800)#246
RDKEMW-20801: widen OCDM auto-start window to ~200s (-r 250 -d 800)#246sergiygladkyy wants to merge 1 commit into
Conversation
There was a problem hiding this comment.
Pull request overview
This PR updates the OCDM systemd oneshot unit to tolerate longer device-provisioning delays by increasing the PluginActivator retry count during service start.
Changes:
- Increased
wpeframework-ocdm.servicePluginActivator retry count from-r 200to-r 400to extend the maximum precondition polling window.
| Type=oneshot | ||
| RemainAfterExit=yes | ||
| ExecStart=/usr/bin/PluginActivator -r 200 OCDM | ||
| ExecStart=/usr/bin/PluginActivator -r 400 OCDM |
There was a problem hiding this comment.
This needs to be tested. Will include in DEV QA flow. We can add a param TimeoutStartSec=360, but by checking the actual logs, we can tell that it's not 90s already, it is more, since -r 200 is 100+s and the unit fails on attempts always and never on timeout
…sioning Reason for change: post-FSR provisioning can outlast OCDM's activation poll, which then gives up and fails the Thunder target. Widen the poll to ~200s via -r 250 -d 800. PluginActivator maxRetries is uint8_t (max 255) - -r 400 wrapped to 144 - so the window is widened with the retry delay (-d), not the count. Follows RDKEMW-5207, which set -r 200 for the same reason. Test Procedure: FSR then connect with provisioning delayed past ~130s; confirm wpeframework-ocdm.service reaches active (exited) instead of failed, and the poll runs the full 250-attempt window. Implements: updates to wpeframework-ocdm.service Risks: No Source: COMCAST License: Apache-2.0 Upstream-Status: Pending Signed-off-by: Sergiy Gladkyy <sgladkyy@productengine.com>
d3a37b7 to
3423ed4
Compare
| [Service] | ||
| Type=oneshot | ||
| RemainAfterExit=yes | ||
| ExecStart=/usr/bin/PluginActivator -r 200 OCDM | ||
| ExecStart=/usr/bin/PluginActivator -r 250 -d 800 OCDM |
There was a problem hiding this comment.
same as before - we will find out this in actual tests. If the result is expected in terms of max time window acc to the params, this is it. If Copilot is right and the window gets cut by sysetmd, we shall consider new systemd params here as suggested
Reason for change
-r 200, ~101s) can expire before the PROVISIONING precondition is met, failing the unit and taking the WPEFramework target down.uint8_t(caps at 255), so the delay is the lever:-r 250 -d 800. (-r 400wraps to 144, shorter than stock.)-r 200the same way.Test Procedure
Described in the ticket: FSR then connect with provisioning delayed past ~130s; confirm
wpeframework-ocdm.servicereachesactive (exited)rather thanfailed.Risks
No. Healthy boots break the poll on first successful activation; the wider window is only used on the slow-provisioning path.