[Nfdump-discuss] -s record and -m are mutually exclusive (was: Re: building 1.6.12 - autoheader)

[Mark D. N. <mn...@wi...>]

On 11/13/2014 9:50 AM, Aleksandar Ciric wrote:
>
> autoreconf, before make, had the same problem myself on centos6
>

This worked great, just wanted to make sure this was recorded for all to see later in
searches :).  For my RPM spec, I did then have to explicitly list the man path as
/usr/share/man, but all was good.

On to the main reason I was trying the new version.  I have a script that is used for
reporting when high data usage is detected and it runs various instances of nfdump during
the analysis and report phase.  During the final step, I get this:

collect_nfdump: /usr/bin/nfdump -o csv -M /usr/local/nfsen/profiles/live/calswcore01 -R
nfcapd.201411140830:nfcapd.201411140930 -n 10 -s srcip/bps -L 200M
collect_nfdump: /usr/bin/nfdump -o csv -M /usr/local/nfsen/profiles/live/calswcore01 -R
nfcapd.201411140830:nfcapd.201411140930 -n 10 -s dstip/bps -L 200M
report_nfdump: /usr/bin/nfdump -q -s record/bytes -o fmt:%sap%dap%pr%byt%bps -O bytes -M
/usr/local/nfsen/profiles/live/calswcore01 -R nfcapd.201411140830:nfcapd.201411140930 -L
1.48G "src host 10.8.25.60"
collect error (-s record and -m are mutually exclusive options)

The first two work, but the last errors out as above.  Note there is no reference to -m in
the command line, just -M, which is completely different.  Looking at nfdump.c, I now see
that print_order is enabled for both -m and -O, so this is actually due to the -O option. 
So....  minor display bug that should be adjusted.

Regards,
Mark