Change in osmo-msc[master]: sms queue: avoid repeated Paging for a failed SMS
gerrit-no-reply at lists.osmocom.org
Mon Mar 11 05:03:15 UTC 2019
Neels Hofmeyr has posted comments on this change. ( https://gerrit.osmocom.org/13173 )
Change subject: sms queue: avoid repeated Paging for a failed SMS
Patch Set 2:
PS1, Line 25: This is bad because an MSC should send a Paging Request exactly once. Retrying
: failed Paging is clearly the task of the BSC, not the MSC
> this is not always true. Do you have a specific reference for this? I know that e.g. […]
I realize the reasoning is a bit flawed here.
If the MSC Pages again, it would Page, wait for timeout, and after that Page again.
So the decision to Page again would be part of the Paging implementation.
If at some point the decision is that Paging expired, the pending transactions should be cleaned up.
What this code is currently doing is Page, then the Paging code decides that it failed.
The SMS transaction gets told that reaching the subscriber has failed.
And then what happens is that the SMS queue starts over, by circumstance picks the very same SMS again,
and then asks the Paging code to start Paging from scratch right away as a consequence of that.
So this is not really about the Paging code, it is the SMS queue code trying to deliver the same failed SMS over and over and over, from a single periodical SMS queue trigger. It is re-triggering Paging from the wrong angle.
So if we also had a Paging retry in osmo-msc, then the SMS queue would restart this series of Paging attempts again and again == it spams Pagings.
I should clarify the commit log...
To view, visit https://gerrit.osmocom.org/13173
To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings
Gerrit-Owner: Neels Hofmeyr <nhofmeyr at sysmocom.de>
Gerrit-Reviewer: Harald Welte <laforge at gnumonks.org>
Gerrit-Reviewer: Jenkins Builder (1000002)
Gerrit-Reviewer: Neels Hofmeyr <nhofmeyr at sysmocom.de>
Gerrit-Comment-Date: Mon, 11 Mar 2019 05:03:15 +0000
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the gerrit-log