You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cassandra.apache.org by Maki Watanabe <wa...@gmail.com> on 2011/04/22 17:38:58 UTC
Concern on HeartBeatState Generation
Hello,
I found Gossipper is initiated with seconds from Unix Epoch
(=System.currentTimeMillis() / 1000)
for HeartBeatState Generation.
Do we get same generation value at very quick restart? Are there any risk here?
regards,
maki
Re: Concern on HeartBeatState Generation
Posted by Watanabe Maki <wa...@gmail.com>.
Thanks!
From iPhone
On 2011/04/23, at 9:31, Jonathan Ellis <jb...@gmail.com> wrote:
> No -- see SystemTable.incrementAndGetGeneration.
>
> On Fri, Apr 22, 2011 at 10:38 AM, Maki Watanabe <wa...@gmail.com> wrote:
>> Hello,
>> I found Gossipper is initiated with seconds from Unix Epoch
>> (=System.currentTimeMillis() / 1000)
>> for HeartBeatState Generation.
>> Do we get same generation value at very quick restart? Are there any risk here?
>>
>> regards,
>>
>> maki
>>
>
>
>
> --
> Jonathan Ellis
> Project Chair, Apache Cassandra
> co-founder of DataStax, the source for professional Cassandra support
> http://www.datastax.com
Re: Concern on HeartBeatState Generation
Posted by Jonathan Ellis <jb...@gmail.com>.
No -- see SystemTable.incrementAndGetGeneration.
On Fri, Apr 22, 2011 at 10:38 AM, Maki Watanabe <wa...@gmail.com> wrote:
> Hello,
> I found Gossipper is initiated with seconds from Unix Epoch
> (=System.currentTimeMillis() / 1000)
> for HeartBeatState Generation.
> Do we get same generation value at very quick restart? Are there any risk here?
>
> regards,
>
> maki
>
--
Jonathan Ellis
Project Chair, Apache Cassandra
co-founder of DataStax, the source for professional Cassandra support
http://www.datastax.com