Experimental Physics and
Industrial Control System

"Zelazny, Michael Stanley" <[email protected]> · Thu, 15 Sep 2011 15:50:04 -0700

>From time to time it appears as though my sequences don't get their monitor callbacks.  Example:

#42 of 196:
Channel name: "KLYS:LI27:51:BVLT"
  Unexpanded (assigned) name: ""
  Variable name: "BVLT"
    address = 203186364 = 0xc1c60bc
    type = float
    count = 1
  Value = 0
  Monitor flag = 0
    Monitored
  Assigned
  Connected
  Get not completed or no get issued
  Put not completed or no put issued
  Status = 0
  Severity = 0
  Message =
  Time stamp = 2011/09/15 14:46:40.940
Next? (+/- skip count)

$ caget KLYS:LI27:51:BVLT KLYS:LI27:51:BVLT.MDEL
KLYS:LI27:51:BVLT              349.717
KLYS:LI27:51:BVLT.MDEL         0

The PV is assigned with pvAssign and monitored with pvMonitor.

The PV lives on the same soft ioc as the sequence.

epics> dbpr("KLYS:LI27:51:BVLT",1000)
ACKS: MAJOR         ACKT: YES           ADEL: 0             ALST: 350.166015625
AOFF: 0             ASG:                ASLO: 0             ASP: 0xa7acfe8
BKPT: 00            DESC: Beam Voltage  DISA: 0             DISP: 0
DISS: NO_ALARM      DISV: 1             DOL:CONSTANT        DPVT: (nil)
DRVH: 0             DRVL: 0             DSET: 0x6bf280      DTYP: Soft Channel
EGU:                EGUF: 0             EGUL: 0             EOFF: 0
ESLO: 1             EVNT: 0             FLNK:DB_LINK KLYS:LI27:51:BVLTTS
HHSV: NO_ALARM      HIGH: 360.000152587891                  HIHI: 0
HOPR: 0             HSV: MAJOR          HYST: 0             INIT: 0
IVOA: Continue normally                 IVOV: 0             LALM: 350.166015625
LBRK: 0             LCNT: 0             LINR: NO CONVERSION LLSV: NO_ALARM
LOLO: 0             LOPR: 0             LOW: 0              LSET: 0xb7b25dc8
LSV: NO_ALARM       MDEL: 0
MLIS: 40 0a 70 8a d0 58 5b 0e 03 00 00 00                   MLOK: c8 b8 f0 09
MLST: 350.166015625 NAME: KLYS:LI27:51:BVLT                 NSEV: NO_ALARM
NSTA: NO_ALARM      OIF: Full           OMOD: 0             OMSL: supervisory
ORAW: 350           ORBV: 0             OROC: 0             OUT:CONSTANT
OVAL: 350.166015625 PACT: 0             PBRK: (nil)         PHAS: 0
PINI: YES           PPN: (nil)          PPNR: (nil)         PREC: 6
PRIO: LOW           PROC: 0             PUTF: 0             PVAL: 350.166015625
RBV: 0              RDES: 0x81bbc50     ROFF: 0             RPRO: 0
RSET: 0xf66840      RVAL: 350           SCAN: Passive       SDIS:CONSTANT
SEVR: NO_ALARM      SIML:CONSTANT       SIMM: NO            SIMS: NO_ALARM
SIOL:CONSTANT       SPVT: (nil)         STAT: NO_ALARM
TIME: 2011-09-15 14:58:57.976294000     TPRO: 0             TSE: 0
TSEL:CONSTANT       UDF: 0              VAL: 350.166015625

More than 99% of the time, the sequence works great, but with 85 klystrons in SLAC's LCLS I see about 1 failure per month.  When the sequence starts one of its PVs appears to never get any callbacks.  If I restart the sequence, everything is fine.

Using EPICS 3.14.12, sequencer 2.0.11, OS RHEL 4, update 8.

Q: Has anyone else seen this?  If so, how did you fix it?

I'm considering upgrading my sequencer to 2.1.2 based on Ben Franksen's notes about fixing bugs in pvAssign and "Safe mode".

Thanks,

Mike

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System