Every thing is possible with Thrift, provided that you manage every thing
manually client side. Having coded an implementation of Achilles (object
mapper) over Hector, it was really painfull to manage collections in Thrift.
Now to stick to the topic, if you want to nest collections into
Is it ok to delete files from backups directory (hardlinks) once I have it
copied over remotely? Any caution to take?
Thanks,Kumar
Despite storing a replica in the backup node, what is the guarantee that the
backup node has all the data? Unless you make consistency a priority over
availability of your cluster.
I could think of another approach.
You can design your cluster with a topology such that your work load is split
Wow.. that's the right answer. Now I'm super excited for C* 2.1 :) ..
yeah.. that would work perfectly. having custom types would perfectly
solve my problem.
Now the issue is whether I wait for the next version or just push through
this version…
On Sat, Jun 14, 2014 at 2:24 AM, DuyHai Doan
Hi everyone,
this week we upgraded one of our Systems from Cassandra 1.2.16 to 2.0.8.
All 3 nodes were upgraded. SStables are upgraded.
Unfortunetaly we are now experiencing that Cassandra starts to hang every
10 hours or so.
We can see the MemoryMeter being very active, every time it is
You should delete the backup files once you have copied them off. Otherwise
they will start to use disk space as the live SSTables diverge from the
snapshots/incrementals.
-psanford
On Sat, Jun 14, 2014 at 10:17 AM, S C as...@outlook.com wrote:
Is it ok to delete files from backups directory
I am thinking of rm file.db once the backup is complete. Any special cases
to be careful about?
-Kumar
Date: Sat, 14 Jun 2014 13:13:10 -0700
Subject: Re: incremental backups
From: psanf...@retailnext.net
To: user@cassandra.apache.org
You should delete the backup files once you have copied them
Hi there,
I was wondering if there is a good reason for select queries on secondary
indexes to not support any where operator other than the equality operator,
or if its just a missing feature in CQL.
Thanks,
Tommaso
