Date: Fri, 27 May 2016 18:02:28 +0200 From: Pierre Baudry pierre.baudry@diateam.net I noticed that since several weeks that osmo-pcu did not work as good as previously.
Yes we have also seen similar failures for GPRS calls in the osmo-pcu. Our setup uses NuRAN hardware with osmo-bts-sysmo
The latest "good" revision I am aware of is commit d87e1d6ab747423d3668c74d16201a5d967accf0 (2015/12/14)
This is a useful information for us that the above mentioned commit works for GPRS.
Can somebody confirm or reproduce this attitude ?
The PCU log that we have during failure matches with the one you have reported. We have tried a workaround for this GPRS call failure by forcing two-phase-access in osmo-pcu.cfg. Based on our analysis done till now it seems One phase access of GPRS is broken.
I will try to provide more logs and a pcap capture as soon as I get hands back on the hardware.
It will be quite useful to narrow down the issue, so please share the relevant logs collected
Regards Saurabh
On Mon, May 30, 2016 at 2:58 AM, Saurabh Sharan Saurabh.Sharan@radisys.com wrote:
Date: Fri, 27 May 2016 18:02:28 +0200 From: Pierre Baudry pierre.baudry@diateam.net I noticed that since several weeks that osmo-pcu did not work as good as previously.
Yes we have also seen similar failures for GPRS calls in the osmo-pcu. Our setup uses NuRAN hardware with osmo-bts-sysmo
I can confirm the same behaviour on B200 using master branches of osmo-pcu, osmo-bts-trx, and osmo-trx.
The latest "good" revision I am aware of is commit d87e1d6ab747423d3668c74d16201a5d967accf0 (2015/12/14)
Confirm.
Can somebody confirm or reproduce this attitude ?
The PCU log that we have during failure matches with the one you have reported. We have tried a workaround for this GPRS call failure by forcing two-phase-access in osmo-pcu.cfg. Based on our analysis done till now it seems One phase access of GPRS is broken.
Confirm matching PCU log and that enabling two-phase-access works around the issue.
-TT
On Wed, Jun 15, 2016 at 06:03:29PM -0700, Tom Tsou wrote:
On Mon, May 30, 2016 at 2:58 AM, Saurabh Sharan Saurabh.Sharan@radisys.com wrote:
Date: Fri, 27 May 2016 18:02:28 +0200 From: Pierre Baudry pierre.baudry@diateam.net I noticed that since several weeks that osmo-pcu did not work as good as previously.
Yes we have also seen similar failures for GPRS calls in the osmo-pcu. Our setup uses NuRAN hardware with osmo-bts-sysmo
I can confirm the same behaviour on B200 using master branches of osmo-pcu, osmo-bts-trx, and osmo-trx.
The latest "good" revision I am aware of is commit d87e1d6ab747423d3668c74d16201a5d967accf0 (2015/12/14)
Confirm.
Maybe related: I have recently found osmo-pcu master to not work for ip.access nanobts (and probably for all BTS models) and traced it to commit f1a7b8f tbf: Add state WAIT_ASSIGN
See https://gerrit.osmocom.org/218
We still have to come to a conclusion on whether we want to merge the revert. If anyone is aware of this, any details would be appreciated.
Thanks!
~Neels
On Thu, Jun 16, 2016 at 4:43 AM, Neels Hofmeyr nhofmeyr@sysmocom.de wrote:
Maybe related: I have recently found osmo-pcu master to not work for ip.access nanobts (and probably for all BTS models) and traced it to commit f1a7b8f tbf: Add state WAIT_ASSIGN
See https://gerrit.osmocom.org/218
We still have to come to a conclusion on whether we want to merge the revert. If anyone is aware of this, any details would be appreciated.
Reverting "f1a7b8f tbf: Add state WAIT_ASSIGN" restores single phase GPRS access on osmo-bts-trx / osmo-trx.
End-to-end functionality appears fine, but I cannot comment on if or how the change affects protocol or performance on a more detailed level.
-TT
-
Neels Hofmeyr -
Saurabh Sharan -
Tom Tsou