Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors

RRC establishment fails

Viewing 15 posts - 1 through 15 (of 21 total)
  • Author
    Posts
  • #69302
    LLme
    Guest

    Hi all,

    I am having problems with few cells where number of unsuccessful RRC establishments is unusally high ( counter: pmTotNoRrcConnectReq, vendor: Ericsson ), but at the same time I don’t have any unsuccessful CS or PS RRC establishments ( counters pmTotNoRrcConnectReqCs and pmTotNoRrcConnectReqPs are both zero ). I haven’t found any other counters to determine which kind of traffic could be having problems during RRC establishment, so if you could help me determine what is causing this fails?

    BR,
    LLme

    #69303
    pix
    Guest

    hi llme,

    pmTotNoRrcConnectReq shows the number of RRC requests, not the failures !!
    Of course pmTotNoRrcConnectReq is very high, it represents all the RRC requests in your cell.

    The failures are shown by the difference between pmTotNoRrcConnectReq and pmTotNoRrcConnectReqSuccess

    You can also find all the subcauses of failures below :
    rrc denied adm
    rrc fail after adm
    utran rejection
    loadsharing request

    regards

    #69304
    LLme
    Guest

    Hi pix, thank you for your answer!

    Yes, you are right, difference between pmTotNoRrcConnectReq and pmTotNoRrcConnectReqSuccess
    is high, not the counter pmTotNoRrcConnectReq alone.

    I have checked counter pmNoFailedAfterAdm and realise that it’s value is almost the same as difference between pmTotNoRrcConnectReq and pmTotNoRrcConnectReqSuccess on that problematic day.

    What are the most common reasons for failed RRCs and RABs after admission?

    #69305
    pix
    Guest

    hi llme,

    Thanks a lot for sharing your issue here, it’s very interesting.

    Failure after Adm are a pain in the a$$…. Did it increase suddenly, or was it always high ?

    Possible causes :
    1/ TN issue (Iub)
    2/ RNC issue (sw issue)
    3/ Core network issue

    Try doing a Iub lock + cold restart of the RBS.

    Also, please show the values for one day, so I can have a better understanding.

    pmxh cell=… afteradm|pmTotNoRrcConnectReq -m 4 -a

    If the RBS Restart did not work, then GPEH traces are the next step.

    So let me recap:
    1/ sudden increase ?
    2/ share pmx results
    3/ do a iub lock + rbs cold restart

    Regards
    pix

    #69306
    LLme
    Guest

    Hi pix,

    unfortunately, Iub lock and restart did not help.

    The increase was sudden, it started 10 days ago, and jumped from 90 to 400 a day.

    Here are the stats:

    Start Time: 2013-11-07 08:15:00 End Time: 2013-11-08 08:00:00
    Object
    UtranCell=3G2120B
    pmNoFailedAfterAdm 227
    pmTotNoRrcConnectReq 57021
    pmTotNoRrcConnectReqCs 6420
    pmTotNoRrcConnectReqCsSucc 6412
    pmTotNoRrcConnectReqPs 12147
    pmTotNoRrcConnectReqPsSucc 12138
    pmTotNoRrcConnectReqSms 9166
    pmTotNoRrcConnectReqSubTr 2255
    pmTotNoRrcConnectReqSuccess 56803

    Regards,
    LL

    #69307
    pix
    Guest

    hi,

    In your output, I can’t see the counter related to the failures. Maybe i’m wrong with the name. “aftAdm” perhaps ?
    It contains the string “RRC” 🙂

    Could you do it again with the counters, including the one i’m talking about ?

    Also include all counters related to:
    RrcDenied
    LoadSharing
    RrcLc
    Irat

    Thanks,

    Cheers
    pix

    #69308
    LLme
    Guest

    That’s the problem, I’m not sure what counters are you talking about exactl:)

    I am calcualting failed RRCs as:
    pmTotNoRrcConnectReq-pmTotNoRrcConnectReqSuccess

    Failed CS RRCs as:
    pmTotNoRrcConnectReqCs-pmTotNoRrcConnectReqCsSucc

    and failed PS RRCs as:
    pmTotNoRrcConnectReqPs-pmTotNoRrcConnectReqPsSucc

    So, for example, newest stats are:

    Start Time: 2013-11-22 05:30:00 End Time: 2013-11-22 09:15:00
    pmTotNoRrcConnectReq 7310
    pmTotNoRrcConnectReqSuccess 6356
    pmTotNoRrcConnectReqCs 770
    pmTotNoRrcConnectReqCsSucc 769
    pmTotNoRrcConnectReqPs 706
    pmTotNoRrcConnectReqPsSucc 704

    which means there are:
    7310-6356=954 failed RRCs,
    770-769=1 failed CS RRC,
    706-704=2 failed PS RRCs.

    Other stats you’ve asked for:

    Object
    pmNoLoadSharingRrcConn 0
    pmNoLoadSharingRrcConnCs 0
    pmNoLoadSharingRrcConnPs 0
    pmTotalTimeRejRrcConnRrcLc 0
    pmNoRejRrcConnRrcLc 0
    pmNoReqDeniedAdm 0
    pmNoRlDeniedAdm 0
    pmNoRrcCsReqDeniedAdm 0
    pmNoRrcPsReqDeniedAdm 0
    pmNoRrcReqDeniedAdm 0
    pmNoRrcReqDeniedAdmDlChnlCode 0
    pmNoRrcReqDeniedAdmDlHw 0
    pmNoRrcReqDeniedAdmDlPwr 0
    pmNoRrcReqDeniedAdmUlHw 0
    pmTotNoRrcConnectAttIratCellResel 1262
    pmTotNoRrcConnectSuccessIratCellResel 1049

    BR,
    LL

    #69309
    pix
    Guest

    hi lily,

    Ok great 🙂 Let me go back to work on monday and i’ll give you the names of the counters that are used to classify “failure after adm”.

    Because there are only 900 fails, which are all happening on RRC “signalling” (not CS, not PS), I strnogly expect a real radio issue. Signalling usually occurs at cell border, where radio conditions are the worst.
    Please ensure you didn’t change any neighbours parameter, utranrelation parameters or that a neighbour is not down.

    On monday, we will perform some kpi collection. the best would be to perform some traces (gpeh), but anyway, i’ll do my best with counters.

    regards

    #69310
    LLme
    Guest

    Thanks pix!

    In the meantime, another cell is also having similar problems. But, it’s degradation started on the day we have increased number of HsPdschCodes to 15. Is there any corelation between this and the number of unsuccessful RRCs for signalling?

    #69311
    pix
    Guest

    llme,

    no, it’s not related.

    In your shoes, i would just do a gpeh trace, extract a failure after adm, and see the cause 🙂 That takes one hour and then you’re done !

    In your previous pmx extract, it’s a pity you didn’t put the number of fail after adm. In a previous post though, you stated it is equal to the number of rrc failures.

    And i told you the main causes for fail after adm : it is a RNC or transport network or Core Network issue.
    Have you chercked your RNC load ? Maybe the RNC is in overload ?

    GPEH would tell you all that. If you have no decoder for gpeh trace, i can help (but take a very small trace : 30mn in one worst site + CM export).

    Cheers

    #69312
    LLme
    Guest

    RNC is over-dimensioned, it shouldn’t be the problem.

    I’ve done a gpeh trace, but I don’t know how to analyze it, what to search for? Ive never used it before. Regarding adimssion, I have only found this event: INTERNAL_ADMISSION_CONTROL_RESPONSE.

    #69313
    pix
    Guest

    llme,

    you can share it somewhere (file share)

    and give me the link there : pix_erlang@yahoo.com

    you must also share your CM export from the oss (common explorer ; click on the RNC where you did th trace ; right click ; export NE ; Radio Only, Compressed Mode)

    You give me the gpeh files + this CM, and i can analyse them, i’ve got a tool.

    Regards
    pix

    #69314
    Suzane
    Guest

    Hi,
    Fow which system you guys are facing this Problem, Pix Please share the Outcome of this study..

    Thanks,Suzi

    #69315
    pix
    Guest

    hi,

    it’s on ercisson system.

    any input from your side?

    we had some issues exchanging the gpeh files, and now the files are good but i have not ime to look at them…

    cheers
    pix

    #69316
    jetlogs
    Guest

    Hi pix what could be the cause of utran rejection.upon reading thr docs about it,it seems to be a little bit broad.and what can be done to reduce this?thanks

Viewing 15 posts - 1 through 15 (of 21 total)
  • The forum ‘Telecom Design’ is closed to new topics and replies.