DRTIMEOUT not working as expected?
Omer wanted an HDR secondary with DRAUTO=2 to wait at least 5 minutes before taking over after the primary dies, and found DRTIMEOUT=300 had no such effect. Replies (Jakkrit, Art Kagel) explained the secondary waits roughly four times DRTIMEOUT, so DRTIMEOUT should be 75 for a 5-minute delay, and that the delay only applies to abnormal disconnections (e.g. pulling the network cable) — after onmode -ky or a crash the secondary knows immediately and fails over at once. Omer still wanted a delay in those cases; Art suggested contacting IBM support, and another poster pointed to Connection Manager or a home-grown equivalent. No definitive solution for the delay-on-crash requirement is recorded.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
Hello,
In an HDR environment, it is our requirement that secondary should wait for
the primary(in case of primary failure/crash/shutdown) for atleast 5 minutes
before changing its mode. We know that this can be achieved manually by
setting the DRAUTO=0 but we are required to do it using DRAUTO=2, so that
secondary automaticalkly becomes online(read/write) after minimum 5 minutes of
wait time interval. Infact, we don't want to do it manually.
For this purpose, we have set the following values of HDR configuration
parameters in onconfig of both primary and secondary server.
IDS Version: 11.50.FC3W1
DRINTERVAL 30
DRTIMEOUT 300
DRAUTO 2
DRIDXAUTO 1
CKPTINTVL 300
But it is not working as per our expectations. According to our knowledge, we
have set the DRTIMEOUT value to 300 second (5 minute), but it is not working.
Please suggest that what other parameter need to be tuned accordingly to
achive this goal.
Please let us know the appropriate values of the these parameters to achive
this goal. Or if it is not possible then any other suggestion to achieve this
functionallity is also welcome.
Thanks.
Omer Saeed Khan
HDR will be unpair when ping timeout reach 4 times of DRTIMEOUT
(approximately).
So if you want 5 minute (300) , you should set DRTIMEOUT = 75
BRGs,
Jakkrit A.
________________________________
From: OMER KHAN <oskhan@i2cinc.com>
To: ids@iiug.org
Sent: Tuesday, March 31, 2009 7:43:15 PM
Subject: DRTIMEOUT not working as expected? [15374]
Hello,
In an HDR environment, it is our requirement that secondary should wait for
the primary(in case of primary failure/crash/shutdown) for atleast 5 minutes
before changing its mode. We know that this can be achieved manually by
setting the DRAUTO=0 but we are required to do it using DRAUTO=2, so that
secondary automaticalkly becomes online(read/write) after minimum 5 minutes of
wait time interval. Infact, we don't want to do it manually.
For this purpose, we have set the following values of HDR configuration
parameters in onconfig of both primary and secondary server.
IDS Version: 11.50.FC3W1
DRINTERVAL 30
DRTIMEOUT 300
DRAUTO 2
DRIDXAUTO 1
CKPTINTVL 300
But it is not working as per our expectations. According to our knowledge, we
have set the DRTIMEOUT value to 300 second (5 minute), but it is not working.
Please suggest that what other parameter need to be tuned accordingly to
achive this goal.
Please let us know the appropriate values of the these parameters to achive
this goal. Or if it is not possible then any other suggestion to achieve this
functionallity is also welcome.
Thanks.
Omer Saeed Khan
*******************************************************************************
Forum Note: Use "Reply" to post a response in the discussion forum.
We are unable to get our required result with provided value (75 second) and as a result secondary server become online (read/write) within few second. We need to force secondary server at least 5 minute to become online (read/write).
For a 5 minute timeout, you have to set DRTIMEOUT to 300 seconds/4 or 75 not
300. IDS waits for four times DRTIMEOUT before switching roles.
Art
Art S. Kagel
Oninit (www.oninit.com)
IIUG Board of Directors (art@iiug.org)
Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Oninit, the IIUG, nor any other organization
with which I am associated either explicitly or implicitly. Neither do
those opinions reflect those of other individuals affiliated with any entity
with which I am affiliated nor those of the entities themselves.
On Tue, Mar 31, 2009 at 8:43 AM, OMER KHAN <oskhan@i2cinc.com> wrote:
> Hello,
>
> In an HDR environment, it is our requirement that secondary should wait for
> the primary(in case of primary failure/crash/shutdown) for atleast 5
> minutes
> before changing its mode. We know that this can be achieved manually by
> setting the DRAUTO=0 but we are required to do it using DRAUTO=2, so that
> secondary automaticalkly becomes online(read/write) after minimum 5 minutes
> of
> wait time interval. Infact, we don't want to do it manually.
>
> For this purpose, we have set the following values of HDR configuration
> parameters in onconfig of both primary and secondary server.
>
> IDS Version: 11.50.FC3W1
>
> DRINTERVAL 30
> DRTIMEOUT 300
> DRAUTO 2
> DRIDXAUTO 1
> CKPTINTVL 300>
> But it is not working as per our expectations. According to our knowledge,
> we
> have set the DRTIMEOUT value to 300 second (5 minute), but it is not
> working.
> Please suggest that what other parameter need to be tuned accordingly to
> achive this goal.
>
> Please let us know the appropriate values of the these parameters to achive
> this goal. Or if it is not possible then any other suggestion to achieve
> this
> functionallity is also welcome.
>
> Thanks.
> Omer Saeed Khan
>
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>
--0016364ef07cd0448b04666b29ed
Wait. How are you testing this? If you test it by performing a nice
shutdown of the primary using onmode -ky etc, then the secondary know
immediately that the primary is offline and takes over immediately. This
feature is for normal maintenance shutdowns. To test the failover timeout,
you have to perform a non-normal disconnection between the two servers. Try
unplugging the network cable from the secondary, that should trigger the 5
minute shutdown that you desire to see. It is also the scenario closest to
a real world cause of a failover.
Art
Art S. Kagel
Oninit (www.oninit.com)
IIUG Board of Directors (art@iiug.org)
Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Oninit, the IIUG, nor any other organization
with which I am associated either explicitly or implicitly. Neither do
those opinions reflect those of other individuals affiliated with any entity
with which I am affiliated nor those of the entities themselves.
On Tue, Mar 31, 2009 at 10:05 AM, OMER KHAN <oskhan@i2cinc.com> wrote:
> We are unable to get our required result with provided value (75 second)
> and
> as a result secondary server become online (read/write) within few second.
> We
> need to force secondary server at least 5 minute to become online
> (read/write).
>
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>
--0016364ee89c6bd1d604666b39a7
Hello Art ,
Thanks for your comment. But we need to know other way to implement this, as
we want that secondary database sever would wait for 5 minute minimum in case
of primary database down either it is force fully down using “onmode
-ky” or server/instance crash
Hello Art ,
Thanks for your comment. But we need to know other way to implement this, as
we want that secondary database sever would wait for 5 minute minimum in case
of primary database down either it is force fully down using onmode -ky or
server/instance crash
we have also tested this scenario using kill command for demonstrating the
instance crash but it gives us the same behavior as in onmode -ky or onmode
-sky.
Contact IBM support for a final answer, but FAIK it works the way that I
described it.
Art
Art S. Kagel
Oninit (www.oninit.com)
IIUG Board of Directors (art@iiug.org)
Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Oninit, the IIUG, nor any other organization
with which I am associated either explicitly or implicitly. Neither do
those opinions reflect those of other individuals affiliated with any entity
with which I am affiliated nor those of the entities themselves.
On Thu, Apr 2, 2009 at 3:34 AM, OMER KHAN <oskhan@i2cinc.com> wrote:
> Hello Art ,
>
> Thanks for your comment. But we need to know other way to implement this,
> as
> we want that secondary database sever would wait for 5 minute minimum in
> case
> of primary database down either it is force fully down using “onmode
> -ky” or server/instance crash
>
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>
--0016e6441e28664e4304669720c7
> -----Original Message-----
> From: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] On Behalf Of
> OMER KHAN
> Sent: Thursday, April 02, 2009 6:59 AM
> To: ids@iiug.org
> Subject: Re: DRTIMEOUT not working as expected? [15403]
>
> Hello Art ,
>
> Thanks for your comment. But we need to know other way to implement
> this, as
> we want that secondary database sever would wait for 5 minute minimum
> in case
> of primary database down either it is force fully down using onmode -ky
> or
> server/instance crash
>
> we have also tested this scenario using kill command for demonstrating
> the
> instance crash but it gives us the same behavior as in onmode -ky or
> onmode
> -sky.>
Omer,
Sound to me like you need to implement something like Connection Manager,
which is only available in version 11.50. Here is my presentation on a "home
grown" type of connection manager for our business-critical HDR pair, which is
still running on v10.00. Maybe it is something you could build and use (you
need to be an IIUG member to get at the presentation -- membership is free)...
http://www.iiug.org/idug06/M10.pdf
HTH,
Paul M.