Cassandra 1.x upgrade to Cassandra 2.x

["Caraballo, Rafael" <ra...@twcable.com>]

Hi,

When planning an upgrade on a 20 node multi DC cluster from Cassandra 1.x to Cassandra 2.x version, does the order of running upgradesstables matter?

What is the difference between:
(a) upgrade Cassandra and run upgradesstables one node at a time;
-versus-
(b) upgrade every Cassandra node, then start the upgradesstables one node at a time?

Thanks,
Rafael


________________________________

This E-mail and any of its attachments may contain Time Warner Cable proprietary information, which is privileged, confidential, or subject to copyright belonging to Time Warner Cable. This E-mail is intended solely for the use of the individual or entity to which it is addressed. If you are not the intended recipient of this E-mail, you are hereby notified that any dissemination, distribution, copying, or action taken in relation to the contents of and attachments to this E-mail is strictly prohibited and may be unlawful. If you have received this E-mail in error, please notify the sender immediately and permanently delete the original and any copy of this E-mail and any printout.

Re: Cassandra 1.x upgrade to Cassandra 2.x

[Robert Coli <rc...@eventbrite.com>]

On Thu, Oct 8, 2015 at 9:40 AM, Caraballo, Rafael <
rafael.caraballo@twcable.com> wrote:

> Is there any reason to wait for all the Cassandra nodes to be upgrades
> from 1.x to 2.x before running the first upgradesstables?
>

Not really, because the new version nodes will start to flush new version
sstables at which point you can't stream from them anyway.

=Rob

RE: Cassandra 1.x upgrade to Cassandra 2.x

["Caraballo, Rafael" <ra...@twcable.com>]

Is there any reason to wait for all the Cassandra nodes to be upgrades from 1.x to 2.x before running the first upgradesstables?


From: Robert Coli [mailto:rcoli@eventbrite.com]
Sent: Thursday, October 08, 2015 11:30 AM
To: user@cassandra.apache.org
Subject: Re: Cassandra 1.x upgrade to Cassandra 2.x

On Thu, Oct 8, 2015 at 7:44 AM, Caraballo, Rafael <ra...@twcable.com>> wrote:
When planning an upgrade on a 20 node multi DC cluster from Cassandra 1.x to Cassandra 2.x version, does the order of running upgradesstables matter?

What is the difference between:
(a) upgrade Cassandra and run upgradesstables one node at a time;
-versus-
(b) upgrade every Cassandra node, then start the upgradesstables one node at a time?

This is becoming a FAQ, I should write a blog post...

Briefly, until upgradesstables is done, one must assume that streaming operations will fail. Streaming is used in bootstrap, decommission and repair. Bootstrap is required to replace or add nodes, decommission is required to remove nodes, and repair is required once every gc_grace_seconds.

One must balance the overhead of running upgradesstables simultaneously on multiple nodes with the duration of exposure to inability-to-use streaming. Often this means running upgradesstables in parallel on all hosts simultaneously.

=Rob


________________________________

This E-mail and any of its attachments may contain Time Warner Cable proprietary information, which is privileged, confidential, or subject to copyright belonging to Time Warner Cable. This E-mail is intended solely for the use of the individual or entity to which it is addressed. If you are not the intended recipient of this E-mail, you are hereby notified that any dissemination, distribution, copying, or action taken in relation to the contents of and attachments to this E-mail is strictly prohibited and may be unlawful. If you have received this E-mail in error, please notify the sender immediately and permanently delete the original and any copy of this E-mail and any printout.

Re: Cassandra 1.x upgrade to Cassandra 2.x

[Robert Coli <rc...@eventbrite.com>]

On Thu, Oct 8, 2015 at 7:44 AM, Caraballo, Rafael <
rafael.caraballo@twcable.com> wrote:

> When planning an upgrade on a 20 node multi DC cluster from Cassandra 1.x
> to Cassandra 2.x version, does the order of running upgradesstables matter?
>
>
>
> What is the difference between:
>
> (a) upgrade Cassandra and run upgradesstables one node at a time;
>
> -versus-
>
> (b) upgrade every Cassandra node, then start the upgradesstables one node
> at a time?
>

This is becoming a FAQ, I should write a blog post...

Briefly, until upgradesstables is done, one must assume that streaming
operations will fail. Streaming is used in bootstrap, decommission and
repair. Bootstrap is required to replace or add nodes, decommission is
required to remove nodes, and repair is required once every
gc_grace_seconds.

One must balance the overhead of running upgradesstables simultaneously on
multiple nodes with the duration of exposure to inability-to-use streaming.
Often this means running upgradesstables in parallel on all hosts
simultaneously.

=Rob