Skip to Content
avatar image
Former Member

IQ 15.4 upgrade to IQ 16.0 N-Bit FP migration with more HW RAM

My current IQ 15.4 ESD 5.3 is about over 20TB data, the HW is HP Linux with 1TB RAM already.

I ran couple times of N-bit FP conversion at UAT / test env; the N-bit conversion can take over 12 hours.

12 hours long does not include HG/LF index rebuild. In the rebuild process, I do notice %sys can run over 10% from UNIX vmstat.

IQ 16 has new feature like RLV, which also uses some memory.

Should I plan upgrade the HW RAM to 2TB down the road?

Can 2TB shorten the N-Bit conversion and add more benefits on IQ 16?

Thanks

Rey

Add comment
10|10000 characters needed characters exceeded

  • Get RSS Feed

1 Answer

  • Jun 09, 2016 at 03:55 PM

    No,  it depend on cardinality and IO subsystem.  You can  tune FP_NBIT_AUTOSIZE_LIMIT,FP_NBIT_ROLLOVER_MAX_MB and FP_NBIT_ENFORCE_LIMITS.

    Add comment
    10|10000 characters needed characters exceeded