Actions
Bug #6433
openDLGLOBAL ERROR counter group 'hnb' already exists for index 0, instead using index 1. This is a software bug that needs fixing. (rate_ctr.c:221)
Status:
New
Priority:
Normal
Assignee:
-
Target version:
-
Start date:
04/03/2024
Due date:
% Done:
0%
Spec Reference:
Description
hnb_persistent_alloc() always passes index 0 to rate_ctr_group_alloc(),
causing the error log on each new HNB attach:
DLGLOBAL ERROR counter group 'hnb' already exists for index 0, instead using index 1. This is a software bug that needs fixing. (rate_ctr.c:221)
Cause is hnbgw.c hnb_persistent_alloc()
hnbp->ctrs = rate_ctr_group_alloc(hnbp, &hnb_ctrg_desc, 0);
There should be a definitive id per each new HNB, so that vty or ctrl commands can use the same ID as used for rate counters,
things like 'bts 23' in osmo-bsc.
This was added in
commit e4d0951a86064f92cc4ff76bbbdc0bc88e2d94d3 Refs: 1.5.0-28-ge4d0951 Author: Harald Welte <laforge@osmocom.org> AuthorDate: Sat Feb 24 11:38:55 2024 +0100 Commit: Harald Welte <laforge@osmocom.org> CommitDate: Tue Mar 12 21:35:40 2024 +0100 New per-hnb rate_ctr and stat_item groups; track uptime Let's add a per-hnb rate_ctr and stat_item group. Only one initial counter (number of Iuh establishments) and one initial stat_item (uptime/downtime) is implemented. Related: SYS#6773 Change-Id: I26d7c3657cdaf7c6ba5aa10a0677381ab099f8dd
Updated by laforge 29 days ago
- Assignee deleted (
laforge)
While we were discussing the counters, I was advised that the modern way to do this is to use the related _set_name functions and not rely on the old integer IDs of prehistoric time . And indeed, the rate counters are exposed with the string names that are specified.
Un-assigning myself as given my current backlog I won't be able to look into this anytime soon.
Actions