Determining the current 'snapmirror throttle' value?

View: New views
5 Messages — Rating Filter:   Alert me  

Determining the current 'snapmirror throttle' value?

by John Stoffel-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message


Guys,

Does anyone know how to determine what the current setting of
'snapmirror throttle N destination' actually is?  I've tried looking
at the output of:

   snapmirror status -l
   snapmirror throttle

We're looking to dynamically throttle and un-throttle connections, and
it would be nice to confirm what the actual setting is without have to
guess and just set the value blindly.

We're running 7.2.6.1 and moving to 7.3.1.1 soon, or something newer.

Thanks,
John
    John Stoffel - Senior Staff Systems Administrator - System LSI Group
  Toshiba America Electronic Components, Inc. - http://www.toshiba.com/taec
             john.stoffel@... - 508-486-1087

RE: Determining the current 'snapmirror throttle' value?

by Leimann, Olaf :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message


Perhaps a silly question, but if "you" are doing the throttle
change, then why are you not logging that yourself for future
reference ?
Script the thing (perl, sh, whatever).

Another thing to look at: divide amount transferred by current transfer
time.


Olaf Leimann
Escalations Engineer
EMEA Escalations

NetApp
00800-44-638277 EMEA Toll-free
1-877-2638277 USA Toll-free
olaf@...
now.netapp.com


-----Original Message-----
From: John Stoffel [mailto:john.stoffel@...]
Sent: maandag 9 november 2009 21:48
To: toasters@...
Subject: Determining the current 'snapmirror throttle' value?


Guys,

Does anyone know how to determine what the current setting of
'snapmirror throttle N destination' actually is?  I've tried looking
at the output of:

   snapmirror status -l
   snapmirror throttle

We're looking to dynamically throttle and un-throttle connections, and
it would be nice to confirm what the actual setting is without have to
guess and just set the value blindly.

We're running 7.2.6.1 and moving to 7.3.1.1 soon, or something newer.

Thanks,
John
    John Stoffel - Senior Staff Systems Administrator - System LSI Group
  Toshiba America Electronic Components, Inc. -
http://www.toshiba.com/taec
             john.stoffel@... - 508-486-1087


RE: Determining the current 'snapmirror throttle' value?

by John Stoffel-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message


Leimann> Perhaps a silly question, but if "you" are doing the throttle
Leimann> change, then why are you not logging that yourself for future
Leimann> reference ?  Script the thing (perl, sh, whatever).

Yeah, it is a silly question.  Let me push back on you, why can't
NetApp *tell* me what the current limit is?  What's so hard about:

       snapmirror throttle -q [<dest>]

or something like that?  My issue is that if sysadmin A makes a
change, then script B makes another change, then sysadmin C want's to
confirm the settings... it's just not possible.  The only current
answer is to just re-do your 'snapmirror throttle N <dest>' commands
to force things to be what you expect.  

Leimann> Another thing to look at: divide amount transferred by
Leimann> current transfer time.

Sure, I guess I could pull that out with a perl script to parse the
snapmirror status -v output and do the math.  But that only gives the
number over the lifetime of that transfer.  And if the throttle will
work on on-going transfers, it doesn't help me because I could have a
multi-day transfer that gets throttled during work hours, then
un-throttled at night, and throttled again the next time.

So the transfer rate will change and the number you propose I use
won't help.

Thanks,
John
    John Stoffel - Senior Staff Systems Administrator - System LSI Group
  Toshiba America Electronic Components, Inc. - http://www.toshiba.com/taec
             john.stoffel@... - 508-486-1087

RE: Determining the current 'snapmirror throttle' value?

by Borders, Rich :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Sounds like an RFE for me.

Richard D Borders
CPR Escalations Engineer
RTP, North Carolina USA - NetApp, Inc.
Email: rborders@...
Phone:(919) 476-5236
Cell: (919) 757 3727


> -----Original Message-----
> From: John Stoffel [mailto:john.stoffel@...]
> Sent: Thursday, November 12, 2009 10:34 AM
> To: Leimann, Olaf
> Cc: John Stoffel; toasters@...
> Subject: RE: Determining the current 'snapmirror throttle' value?
>
>
> Leimann> Perhaps a silly question, but if "you" are doing the throttle
> Leimann> change, then why are you not logging that yourself for future
> Leimann> reference ?  Script the thing (perl, sh, whatever).
>
> Yeah, it is a silly question.  Let me push back on you, why can't
> NetApp *tell* me what the current limit is?  What's so hard about:
>
>        snapmirror throttle -q [<dest>]
>
> or something like that?  My issue is that if sysadmin A makes a
> change, then script B makes another change, then sysadmin C want's to
> confirm the settings... it's just not possible.  The only current
> answer is to just re-do your 'snapmirror throttle N <dest>' commands
> to force things to be what you expect.
>
> Leimann> Another thing to look at: divide amount transferred by
> Leimann> current transfer time.
>
> Sure, I guess I could pull that out with a perl script to parse the
> snapmirror status -v output and do the math.  But that only gives the
> number over the lifetime of that transfer.  And if the throttle will
> work on on-going transfers, it doesn't help me because I could have a
> multi-day transfer that gets throttled during work hours, then
> un-throttled at night, and throttled again the next time.
>
> So the transfer rate will change and the number you propose I use
> won't help.
>
> Thanks,
> John
>     John Stoffel - Senior Staff Systems Administrator - System LSI
> Group
>   Toshiba America Electronic Components, Inc. -
> http://www.toshiba.com/taec
>     john.stoffel@... - 508-486-1087


RE: Determining the current 'snapmirror throttle' value?

by Leimann, Olaf :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message


You could open a case and ask for it to be linked to "Bug" number
315117.

It has only one customer attached to it yet (not TEAC), and as
an RFE may not see the light for a while (unless you're willing
to pay for it through Sales: a PVR).
Also, it's not public yet, but I'll give you the header info here:

id                      315117
title                   RFE for the ability to issue command to see
current snapmirror throttle setting
state                   NEW

Current target ONTAP is 8.1, but not confirmed.

HOWEVER...

There is also a global throttle option since 7.2RC1:
<http://now.netapp.com/NOW/knowledge/docs/ontap/rel727/html/ontap/cmdref
/man1/na_options.1.htm>
-------------------------
replication.throttle.enable

Enables global network throttling of snapmirror and snapvault transfers.
The default value for this options is off.
-------------------------

This means you can schedule your bandwidth for all SnapMirror/SnapVault
transfers in one go.


Cheers,
Olaf Leimann
Escalations Engineer
EMEA Escalations

NetApp
00800-44-638277 EMEA Toll-free
1-877-2638277 USA Toll-free
olaf@...
now.netapp.com


-----Original Message-----
From: John Stoffel [mailto:john.stoffel@...]
Sent: donderdag 12 november 2009 16:34
To: Leimann, Olaf
Cc: John Stoffel; toasters@...
Subject: RE: Determining the current 'snapmirror throttle' value?


Leimann> Perhaps a silly question, but if "you" are doing the throttle
Leimann> change, then why are you not logging that yourself for future
Leimann> reference ?  Script the thing (perl, sh, whatever).

Yeah, it is a silly question.  Let me push back on you, why can't
NetApp *tell* me what the current limit is?  What's so hard about:

       snapmirror throttle -q [<dest>]

or something like that?  My issue is that if sysadmin A makes a
change, then script B makes another change, then sysadmin C want's to
confirm the settings... it's just not possible.  The only current
answer is to just re-do your 'snapmirror throttle N <dest>' commands
to force things to be what you expect.  

Leimann> Another thing to look at: divide amount transferred by
Leimann> current transfer time.

Sure, I guess I could pull that out with a perl script to parse the
snapmirror status -v output and do the math.  But that only gives the
number over the lifetime of that transfer.  And if the throttle will
work on on-going transfers, it doesn't help me because I could have a
multi-day transfer that gets throttled during work hours, then
un-throttled at night, and throttled again the next time.

So the transfer rate will change and the number you propose I use
won't help.

Thanks,
John
    John Stoffel - Senior Staff Systems Administrator - System LSI Group
  Toshiba America Electronic Components, Inc. -
http://www.toshiba.com/taec
             john.stoffel@... - 508-486-1087