ue, Apr 5, 2016 at 9:34 AM, Jonathan Haddad
>> wrote:
>>
>>> Why is this a requirement? Honestly I don't know why you would do this.
>>>
>>>
>>> On Sat, Apr 2, 2016 at 8:06 PM Mukil Kesavan
>>> wrote:
>>>
>>>> Hell
Hello,
We currently have 3 Cassandra servers running in a single datacenter with a
replication factor of 3 for our keyspace. We also use the SimpleSnitch
wiith DynamicSnitching enabled by default. Our load balancing policy is
TokenAwareLoadBalancingPolicy with RoundRobinPolicy as the child. This
o
a low stratum NTP server.
>
> -- Brice
>
>
>
>
> On Thu, Mar 31, 2016 at 10:00 AM -0700, "Eric Evans" > wrote:
>
> On Wed, Mar 30, 2016 at 8:07 PM, Mukil Kesavan wrote:
>> > Are there any issues if this causes a huge time correction on the cassand
Hi,
We run a 3 server cassandra cluster that is initially NTP synced to a
single physical server over LAN. This server does not have connectivity to
the internet for a few hours to sometimes even days. In this state we
perform some schema operations and reads/writes with QUORUM consistency.
Later
Hello,
The documentation on atomic batches says:
"For example, there is no batch isolation. Clients are able to read the
first updated rows from the batch, while other rows are still being updated
on the server."
However does a write-write conflict scenario arise when I'm updating a
column in tw
