You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cassandra.apache.org by shimi <sh...@gmail.com> on 2010/07/14 17:29:09 UTC
get_range_slices return the same rows
I wrote a code that iterate on all the rows by using get_range_slices.
for the first call I use KeyRange from "" to "".
for all the others I use from <the last key that I got in the previous
iteration> to "".
I always get the same rows that I got in the previous iteration. I tried
changing the batch size but I still gets the same results.
I tried it both in single node and a cluster.
I use RP with version 0.6.3 and Hector.
Does anyone know how this can be done?
Shimi
Re: Seed and nodetool
Posted by Aaron Morton <aa...@thelastpickle.com>.
Can you do an insert with CL ALL?
Are there any ERRORs in the log file? Try turning the logging up the TRACE and see whats happening.
Check B and see A by ssh'ing into B and using node tool from there to connect to A.
Do you have any switches / firewalls between the nodes ? Could this be happening http://www.mail-archive.com/user@cassandra.apacheorg/msg03813.html
Good luck
Aaron
On 15 Jul, 2010,at 03:55 PM, Claire Chang <cl...@merchantcircle.com> wrote:
> BTW,
>
> A is 192.168.11.29
> B is 192.168.11.28
> C is 192.168.11.27
>
> from the result of nodetool ring, does it mean that B thinks A, C are down and C thinks B is down?
>
> I tried to restart B and for a bring moment, I didn't get this problem (all the nodes are all from nodetool) but after a while, this problem came back.
>
> What could be the issue?
>
> thanks,
> Claire
>
> On Jul 14, 2010, at 7:22 PM, Claire Chang wrote:
>
> > I have 3 nodes A, B, C with RF=3. When I configure the cluster and before start taking any read/write request, I first start A, put A itself as seed (following in the instructions on wiki), and then start B (put A as the seed) and then start C (also put A as the seed).
> >
> > B and C seem joining the ring correctly and the cluster is working properly but if I run nodetool
> >
> > claire@A:$ nodetool -h A -p 9090 ring
> > Address Status Load Range Ring
> > 170141183460469231731687303715884105726
> > 192.168.11.29 Up 2.29 GB 56713727820156410577229101238628035242 |<--|
> > 192.168.11.28 Up 2.18 GB 113427455640312821154458202477256070484 | |
> > 192.168.11.27 Up 2.29 GB 170141183460469231731687303715884105726 |-->|
> > claire@A:$ nodetool -h B -p 9090 ring
> > Address Status Load Range Ring
> > 170141183460469231731687303715884105726
> > 192.168.11.29 Down 2.28 GB 56713727820156410577229101238628035242 |<--|
> > 192.168.11.28 Up 2.18 GB 113427455640312821154458202477256070484 | |
> > 192.168.11.27 Down 2.28 GB 170141183460469231731687303715884105726 |-->|
> > claire@A:$ nodetool -h C -p 9090 ring
> > Address Status Load Range Ring
> > 170141183460469231731687303715884105726
> > 192.168.11.29 Up 2.29 GB 56713727820156410577229101238628035242 |<--|
> > 192.168.11.28 Down 2.18 GB 113427455640312821154458202477256070484 | |
> > 192.168.11.27 Up 2.29 GB 170141183460469231731687303715884105726 |-->|
> >
> >
> > Any reason why nodetool thinks that some servers are down if pointing to B or C? If my cluster setup correct?
> >
> > thanks,
> > Claire
> >
>
Re: Seed and nodetool
Posted by Claire Chang <cl...@merchantcircle.com>.
BTW,
A is 192.168.11.29
B is 192.168.11.28
C is 192.168.11.27
from the result of nodetool ring, does it mean that B thinks A, C are down and C thinks B is down?
I tried to restart B and for a bring moment, I didn't get this problem (all the nodes are all from nodetool) but after a while, this problem came back.
What could be the issue?
thanks,
Claire
On Jul 14, 2010, at 7:22 PM, Claire Chang wrote:
> I have 3 nodes A, B, C with RF=3. When I configure the cluster and before start taking any read/write request, I first start A, put A itself as seed (following in the instructions on wiki), and then start B (put A as the seed) and then start C (also put A as the seed).
>
> B and C seem joining the ring correctly and the cluster is working properly but if I run nodetool
>
> claire@A:$ nodetool -h A -p 9090 ring
> Address Status Load Range Ring
> 170141183460469231731687303715884105726
> 192.168.11.29 Up 2.29 GB 56713727820156410577229101238628035242 |<--|
> 192.168.11.28 Up 2.18 GB 113427455640312821154458202477256070484 | |
> 192.168.11.27 Up 2.29 GB 170141183460469231731687303715884105726 |-->|
> claire@A:$ nodetool -h B -p 9090 ring
> Address Status Load Range Ring
> 170141183460469231731687303715884105726
> 192.168.11.29 Down 2.28 GB 56713727820156410577229101238628035242 |<--|
> 192.168.11.28 Up 2.18 GB 113427455640312821154458202477256070484 | |
> 192.168.11.27 Down 2.28 GB 170141183460469231731687303715884105726 |-->|
> claire@A:$ nodetool -h C -p 9090 ring
> Address Status Load Range Ring
> 170141183460469231731687303715884105726
> 192.168.11.29 Up 2.29 GB 56713727820156410577229101238628035242 |<--|
> 192.168.11.28 Down 2.18 GB 113427455640312821154458202477256070484 | |
> 192.168.11.27 Up 2.29 GB 170141183460469231731687303715884105726 |-->|
>
>
> Any reason why nodetool thinks that some servers are down if pointing to B or C? If my cluster setup correct?
>
> thanks,
> Claire
>
Seed and nodetool
Posted by Claire Chang <cl...@merchantcircle.com>.
I have 3 nodes A, B, C with RF=3. When I configure the cluster and before start taking any read/write request, I first start A, put A itself as seed (following in the instructions on wiki), and then start B (put A as the seed) and then start C (also put A as the seed).
B and C seem joining the ring correctly and the cluster is working properly but if I run nodetool
claire@A:$ nodetool -h A -p 9090 ring
Address Status Load Range Ring
170141183460469231731687303715884105726
192.168.11.29 Up 2.29 GB 56713727820156410577229101238628035242 |<--|
192.168.11.28 Up 2.18 GB 113427455640312821154458202477256070484 | |
192.168.11.27 Up 2.29 GB 170141183460469231731687303715884105726 |-->|
claire@A:$ nodetool -h B -p 9090 ring
Address Status Load Range Ring
170141183460469231731687303715884105726
192.168.11.29 Down 2.28 GB 56713727820156410577229101238628035242 |<--|
192.168.11.28 Up 2.18 GB 113427455640312821154458202477256070484 | |
192.168.11.27 Down 2.28 GB 170141183460469231731687303715884105726 |-->|
claire@A:$ nodetool -h C -p 9090 ring
Address Status Load Range Ring
170141183460469231731687303715884105726
192.168.11.29 Up 2.29 GB 56713727820156410577229101238628035242 |<--|
192.168.11.28 Down 2.18 GB 113427455640312821154458202477256070484 | |
192.168.11.27 Up 2.29 GB 170141183460469231731687303715884105726 |-->|
Any reason why nodetool thinks that some servers are down if pointing to B or C? If my cluster setup correct?
thanks,
Claire
Re: get_range_slices return the same rows
Posted by Jonathan Ellis <jb...@gmail.com>.
This is a bug. If you can give us data to reproduce with we can fix it faster.
On Wed, Jul 14, 2010 at 10:29 AM, shimi <sh...@gmail.com> wrote:
> I wrote a code that iterate on all the rows by using get_range_slices.
> for the first call I use KeyRange from "" to "".
> for all the others I use from <the last key that I got in the previous
> iteration> to "".
> I always get the same rows that I got in the previous iteration. I tried
> changing the batch size but I still gets the same results.
> I tried it both in single node and a cluster.
> I use RP with version 0.6.3 and Hector.
>
> Does anyone know how this can be done?
>
> Shimi
>
--
Jonathan Ellis
Project Chair, Apache Cassandra
co-founder of Riptano, the source for professional Cassandra support
http://riptano.com