This post is more than 5 years old
295 Posts
0
2984
nx4 - need test my primary standby datamover configuration step by step
before going in production enviroment, i need to test my primary standby configuracion,
doing this in an controlled event, step by step, to check if all components are working,
i want to simulate, a fail in datamover2,
to activate datamover3(standby mode) to primary mode,
then return the good state of datamover2, to a primary mode again,
and deactivate the datamover3 and return it to its original mode (standbymode)
how much time does this takes.
and what steps to follows
what is the impact, with all the share folders,
there will be any disrruptions?
any one had do this. what do a need to know it, before doing this,
please any idea i will be apreciate
Maynor
gbarretoxx1
366 Posts
0
July 7th, 2010 14:00
Hi,
to test the blade ( aka Data Mover ) failover, simply run the following command :
# server_standby server_2 -a mover
To fail it back ( or restore ) to slot_2, use the following command :
# server_standby server_2 -r mover
This operation can usually take from 30 seconds up to 4 minutes, depending on the size of the environment ( # of filesystems, # of checkpoins, # of tree quotas ) and the DART version - the latest versions have some improvements that minimize this times.
Make sure the network infrastructure is setup properly on both Data Movers.
Regarding the impact, it depends on the environment also.
In a NFS environment, if the clients are using 'hard' mount option, you should not see any impact.
In a CIFS environment, the connections will be closed, and reopened after the activity.
In a iSCSI environment, it depends on the initiators configuration. There are some timeout settings that can make the failover/failback operations transparent.
Once you are not on production yet, this is the perfect time to test it, and see what's the impact on your environment.
Regards,
Gustavo Barreto.
Rainer_EMC
8.6K Posts
0
July 7th, 2010 16:00
is your 2nd data mover a standby?
jelucho
295 Posts
0
July 7th, 2010 16:00
Gustavo, thank you for your fast response,
i will try to do my test, and see what happend,
i see that you are an expert in nas storage due your knowlege
i have some question, about NTP time settings on my 2 datamovers
my control stations is taking the time from a NTP server1
and my two datamovers points to the same NTP server1 but the time is two hours more than the NTP servers,
how to fix this.
I am in Guatemala central America (-6 UTC)
in the timezone setings, i try setting -6, and i try setting america/guatemala too
and the same result.
do you have any idea if there is some wrong
thank you
Maynor
gbarretoxx1
366 Posts
0
July 7th, 2010 19:00
Hi,
please take a look at these two solutions :
https://community.emc.com/docs/DOC-7369
https://community.emc.com/docs/DOC-7131
Also, as Rainer pointed, this is valid only for primary data movers. Standby DM's does not have configurations. It inherits all configurations from the primary one.
Regards,
Gustavo Barreto.
sebbyr
99 Posts
1
July 8th, 2010 08:00
An easy way to update/change the timezone on the DMs is to log into the Celerra manager as root. Go to the properties of the datamovers, and you will be able to type in the city most closely associated with your timezone.
Also, as far as failing over the DM to the standby, it is always good practice to first run a verify switch with the server_standy command. Example: server_standby server_2 -v mover or server_standby ALL -v mover
- Sebby Robles
* If this post was able to help or resolve your question, please take a moment to rate it. Thank you. *
jelucho
295 Posts
0
July 8th, 2010 10:00
yes server_3 is datamover in standby mode
so this means that with this config, i dont need to set time to my standby datamover, only to the primary?
i follow the video instruction with the same result,
i have try with this comand too, "server_date server_2 timesvc start ntp x.x.x.x" same result,
then to see if command is sucessfully then a try "server_date server_2 timesvc" and the same
the time is two hours ahead of NTP SERVER hours, what´s wrong?
thanks
Maynor
nandas
1.5K Posts
0
July 8th, 2010 12:00
You are right - for standby data mover, no need to set the NTP (in fact you can not set any NTP server for standby mover) - you may manually set the Timezone and date & time on the standby data mover - but that is not needed, when the active data mover fails over to the standby one - the configuration from Active data mover is used.
In other words, the virtual server name "server_2" remains the same on the standby data mover when it fails. In normal case, server_2 is on slot 2 (data mover2) and server_3 on slot 3 (Data Mover 3) which is the standby - in case of failover, data mover on slot 3 gets renamed as server_2 and continue all tasks as the active data mover.
For the issue you are having with 2 hrs difference in time - I 'll check whether the time zone has been set correctly on the data mover. You may login into the Celerra Manager and go to the property page of the data mover to check or modify the Time Zone for the data mover.
On the CLI - the following command will also show the current time with the time zone
$ server_date server_2
To check whether the NTP server is working properly or not - you may run the following command -
$ server_date server_2 timesvc stats ntp
check for the hits or misses.
If you don't find any hit - that means the NTP server is not reachable from the data mover - run the command
$ server_ping server_2
My 2 cents
Sandip
gbarretoxx1
366 Posts
0
July 8th, 2010 13:00
Hi,
Try to set the Control Station timezone :
/nbsnas/http/webui/bin/timezone.pl -s America/Guatemala
Gustavo Barreto.