iedemam has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bsc/+/27081 )
Change subject: stats: new trackers for lchan life duration
......................................................................
Patch Set 22:
(1 comment)
Patchset:
PS22:
I am starting on this now. I will rework the patch to have one timer per BTS instead of one timer per lchan. My instinct to use one per BTS instead of one global for the entire BSC is so that stat summation activity is spread out in smaller chunks instead of in one large task for potentially thousands of lchans.
--
To view, visit https://gerrit.osmocom.org/c/osmo-bsc/+/27081
To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings
Gerrit-Project: osmo-bsc
Gerrit-Branch: master
Gerrit-Change-Id: I1b0670c47cb5e0b7776eda89d1e71545ba0e3347
Gerrit-Change-Number: 27081
Gerrit-PatchSet: 22
Gerrit-Owner: iedemam <michael(a)kapsulate.com>
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: laforge <laforge(a)osmocom.org>
Gerrit-Reviewer: neels <nhofmeyr(a)sysmocom.de>
Gerrit-Reviewer: pespin <pespin(a)sysmocom.de>
Gerrit-Comment-Date: Thu, 12 May 2022 07:32:25 +0000
Gerrit-HasComments: Yes
Gerrit-Has-Labels: No
Gerrit-MessageType: comment
pespin has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bsc/+/27081 )
Change subject: stats: new trackers for lchan life duration
......................................................................
Patch Set 22:
(1 comment)
Patchset:
PS22:
> I agree. […]
I agree with Harald the main point is avoid having so many timers. So I'd start by migrating to having 1 timer per BTS, or 1 main BSC timer which triggers ever 500ms-1s and which then iterates to update the data for all its depending lchans.
Counting for SDCCH on top of TCH shouldn't make a big different imho if calculated in the same iteration.
--
To view, visit https://gerrit.osmocom.org/c/osmo-bsc/+/27081
To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings
Gerrit-Project: osmo-bsc
Gerrit-Branch: master
Gerrit-Change-Id: I1b0670c47cb5e0b7776eda89d1e71545ba0e3347
Gerrit-Change-Number: 27081
Gerrit-PatchSet: 22
Gerrit-Owner: iedemam <michael(a)kapsulate.com>
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: laforge <laforge(a)osmocom.org>
Gerrit-Reviewer: neels <nhofmeyr(a)sysmocom.de>
Gerrit-Reviewer: pespin <pespin(a)sysmocom.de>
Gerrit-Comment-Date: Wed, 11 May 2022 17:12:07 +0000
Gerrit-HasComments: Yes
Gerrit-Has-Labels: No
Comment-In-Reply-To: iedemam <michael(a)kapsulate.com>
Gerrit-MessageType: comment
iedemam has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bsc/+/27081 )
Change subject: stats: new trackers for lchan life duration
......................................................................
Patch Set 22:
(1 comment)
Patchset:
PS22:
I agree. Some things which would still keep this feature useful to us but will reduce the performance impact:
1) reduce granularity from 0.1s to 1s: While I'd like to keep 100ms resolution, maybe there's no way to do it efficiently.
2) only start timers and track stats on TCH lchans: We're primarily using this to gather Erlang stats and secondarily to glance site health. Even if it's only TCH it still serves that purpose.
The biggest BSC we're talking about has 4480 SDCCH lchans defined and 3200 TCH. The combination of 1 and 2 above gives us a full order of magnitude drop and then cuts that in half. Maybe a better starting point?
--
To view, visit https://gerrit.osmocom.org/c/osmo-bsc/+/27081
To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings
Gerrit-Project: osmo-bsc
Gerrit-Branch: master
Gerrit-Change-Id: I1b0670c47cb5e0b7776eda89d1e71545ba0e3347
Gerrit-Change-Number: 27081
Gerrit-PatchSet: 22
Gerrit-Owner: iedemam <michael(a)kapsulate.com>
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: laforge <laforge(a)osmocom.org>
Gerrit-Reviewer: neels <nhofmeyr(a)sysmocom.de>
Gerrit-Reviewer: pespin <pespin(a)sysmocom.de>
Gerrit-Comment-Date: Wed, 11 May 2022 17:06:11 +0000
Gerrit-HasComments: Yes
Gerrit-Has-Labels: No
Gerrit-MessageType: comment
Attention is currently required from: iedemam.
pespin has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bsc/+/27081 )
Change subject: stats: new trackers for lchan life duration
......................................................................
Patch Set 22:
(1 comment)
Patchset:
PS22:
> So due to severe performance regressions, I think we will have to re-open the discussion about this […]
I also think having subsecond granularity may not be that interesting given the amount of extra performance penalty induced. Maybe if at all have it at 500ms minimum.
--
To view, visit https://gerrit.osmocom.org/c/osmo-bsc/+/27081
To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings
Gerrit-Project: osmo-bsc
Gerrit-Branch: master
Gerrit-Change-Id: I1b0670c47cb5e0b7776eda89d1e71545ba0e3347
Gerrit-Change-Number: 27081
Gerrit-PatchSet: 22
Gerrit-Owner: iedemam <michael(a)kapsulate.com>
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: laforge <laforge(a)osmocom.org>
Gerrit-Reviewer: neels <nhofmeyr(a)sysmocom.de>
Gerrit-Reviewer: pespin <pespin(a)sysmocom.de>
Gerrit-Attention: iedemam <michael(a)kapsulate.com>
Gerrit-Comment-Date: Wed, 11 May 2022 17:00:48 +0000
Gerrit-HasComments: Yes
Gerrit-Has-Labels: No
Comment-In-Reply-To: laforge <laforge(a)osmocom.org>
Gerrit-MessageType: comment
laforge has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bsc/+/27081 )
Change subject: stats: new trackers for lchan life duration
......................................................................
Patch Set 22:
(1 comment)
Patchset:
PS22:
So due to severe performance regressions, I think we will have to re-open the discussion about this code.
It adds one osmo_timer for each lchan, and that timer expires every 100ms. So thinking of a system with 80 BTS and each BTS having a single trx with ~ 16 lchan, we end up with 12800 (!) timers expiring each second! This is certainly the single biggest performance hog in osmo-bsc for any setup larger than the most trivial one.
Whatever we do, we have to do better than that. Like one timer expiring every 100ms iteratinng over the lchans?
--
To view, visit https://gerrit.osmocom.org/c/osmo-bsc/+/27081
To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings
Gerrit-Project: osmo-bsc
Gerrit-Branch: master
Gerrit-Change-Id: I1b0670c47cb5e0b7776eda89d1e71545ba0e3347
Gerrit-Change-Number: 27081
Gerrit-PatchSet: 22
Gerrit-Owner: iedemam <michael(a)kapsulate.com>
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: laforge <laforge(a)osmocom.org>
Gerrit-Reviewer: neels <nhofmeyr(a)sysmocom.de>
Gerrit-Reviewer: pespin <pespin(a)sysmocom.de>
Gerrit-Comment-Date: Wed, 11 May 2022 16:39:27 +0000
Gerrit-HasComments: Yes
Gerrit-Has-Labels: No
Gerrit-MessageType: comment