[Orca-users] Displaying Statistics for NetApp SAN disks in Orca
Ayotunde Itayemi
aitayemi at gmail.com
Wed Jan 21 10:19:39 PST 2009
Thanks David,
I am testing this right now. Can I assume that if I use any of the 2 regxp
you suggested below, it won't "break" the display of other "regular" disk
devices and SVM meta-devices?
Thanks.
2009/1/20 David Michaels <dragon at raytheon.com>
>
> I have got the line below in my /usr/local/lib/orcallator.cfg file. How
> do I modify this to show my external LUNs in addition to internal drives?
>
> In particular the NetApp LUNs.
>
> disk_runp_((?:c\d+t\d+d\d+)|(?:c\d+d\d+)|(?:[ms]d\d+))
>
>
> 4. c13t500A098387393451d0 <NETAPP-LUN-0.2-1.95TB>
>
>
>
> That's a really awesome disk identifier. The regexp above won't work,
> because it's expecting the target to be a number, and in your case it's
> hexadecimal.
>
> First, try using \x after the 't':
>
> disk_runp_((?:c\d+t\x+d\d+)|(?:c\d+d\d+)|(?:[ms]d\d+))
> ^^
>
> If that doesn't work, try the more verbose method:
> disk_runp_((?:c\d+t[0-9A-F]+d\d+)|(?:c\d+d\d+)|(?:[ms]d\d+))
>
>
> A related question:
>
> The server is running Oracle database and some of the LUNs are presented
> directly to the Oracle database as RAW devices.
> How can I get statistics for "Maximum and Average Disk Busy" – if asked?
>
>
>
> If the graphs and datablocks in them are insufficient, you can query the
> raw data directly using orcallator-column -m -c disk_runp_<whatever disk you
> want> <files to scan>
>
> For instance:
>
> me at myserver <~orca/bin><29> *./orcallator_column -m -c disk_runp_c4t45d3
> ../var/orca/orcallator/npd-blizzard/orcallator-2009-01-**
>
> Machine
> ../var/orca/orcallator/npd-blizzard/orcallator-2009-01-01-000.bz2
> .... [ similar garbage lines from bzip'ed files trimmed] ....
>
> Machine locltime
> disk_runp_c4t45d3
> ../var/orca/orcallator/npd-blizzard/orcallator-2009-01-16-004
> 17:20:00 1.73
>
> Machine locltime
> disk_runp_c4t45d3
> ../var/orca/orcallator/npd-blizzard/orcallator-2009-01-20-002
> 14:30:00 0.00
>
>
> If, like me, most of your data files are bzip'ed, you can do something like
> this (note the trailing '-' by itself to orcallator_column, telling it to
> scan standard input instead of scanning particular files).
>
> me at myserver <~/orca/bin><33> *bzcat
> ../var/orca/orcallator/npd-blizzard/orcallator-2009-01-* |
> ./orcallator_column -m -c disk_runp_c4t45d3 -c timestamp -*
> bzcat: Input file name
> ../var/orca/orcallator/npd-blizzard/orcallator-2009-01-16-004 doesn't end in
> `.bz2', skipping.
> bzcat: Input file name
> ../var/orca/orcallator/npd-blizzard/orcallator-2009-01-20-002 doesn't end in
> `.bz2', skipping.
> Machine locltime disk_runp_c4t45d3
> - 10:00:00 94.94
>
> This won't tell you when it happened, though.
>
> Cheers,
> --Dragon
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: </pipermail/orca-users/attachments/20090121/f579a862/attachment.html>
More information about the Orca-users
mailing list