nodetool drain

[tzach]

Drains the node from live traffic
http://docs.datastax.com/en/cassandra/2.1/cassandra/tools/toolsNodetool_r.html

Example usage:
http://devblog.michalski.im/2012/11/25/safe-cassandra-shutdown-and-restart/
http://www.slideshare.net/nmilford/cassandra-for-sysadmins (#32)

[tzach]

@slivne drain is for GA, and not implemented yet

[penberg]

@tzach @slivne, really? I thought 39af3ec added nodetool drain support?

Ping @pdziepak

[pdziepak]

It should be available at REST API level. Perhaps some glue code in scylla-jmx is missing.

[penberg]

StorageService.drain() seems to have the necessary glue. @tzach, hmm?

[tzach]

I tested with 0.13.2 and its not there

nodetool drain
error: API call is not supported yet

[amnonh]

assign it to me

[penberg]

@tzach It's in master, not in 0.13.2.

[tzach]

I can't keep up with you youngsters, I only test RPMs
For 0.13.2 the HTTP is not avilable either

curl -X POST --header "Content-Type: application/json" --header "Accept: application/json" "http://localhost:10000/storage_service/drain"
{"message": "API call is not supported yet", "code": 500}

[amnonh]

I've just check it on head and it works.
So wait for the next release and try again

On Tue, Dec 29, 2015 at 5:24 PM, Tzach Livyatan notifications@github.com
wrote:

I can't keep up with you youngsters, I only test RPMs
For 0.13.2 the HTTP is not avilable either

curl -X POST --header "Content-Type: application/json" --header "Accept: application/json" "http://localhost:10000/storage_service/drain"
{"message": "API call is not supported yet", "code": 500}

—
Reply to this email directly or view it on GitHub
#442 (comment).

[slivne]

there is still a discussion between gleb and pawel (unless they sorted it
out) on impl items in drain.

On Tue, Dec 29, 2015 at 5:30 PM, Amnon Heiman notifications@github.com
wrote:

I've just check it on head and it works.
So wait for the next release and try again

On Tue, Dec 29, 2015 at 5:24 PM, Tzach Livyatan notifications@github.com
wrote:

I can't keep up with you youngsters, I only test RPMs
For 0.13.2 the HTTP is not avilable either

curl -X POST --header "Content-Type: application/json" --header "Accept:
application/json" "http://localhost:10000/storage_service/drain"
{"message": "API call is not supported yet", "code": 500}

—
Reply to this email directly or view it on GitHub
#442 (comment).

—
Reply to this email directly or view it on GitHub
#442 (comment).

[amnonh]

I'm not sure how well it works, but you don't get the experience Tzach was
getting.

Now the drain command completed without any errors.
And scylla prints to the log that it runs drain

On Tue, Dec 29, 2015 at 5:33 PM, slivne notifications@github.com wrote:

there is still a discussion between gleb and pawel (unless they sorted it
out) on impl items in drain.

On Tue, Dec 29, 2015 at 5:30 PM, Amnon Heiman notifications@github.com
wrote:

I've just check it on head and it works.
So wait for the next release and try again

On Tue, Dec 29, 2015 at 5:24 PM, Tzach Livyatan <
notifications@github.com>
wrote:

I can't keep up with you youngsters, I only test RPMs
For 0.13.2 the HTTP is not avilable either

curl -X POST --header "Content-Type: application/json" --header
"Accept:
application/json" "http://localhost:10000/storage_service/drain"
{"message": "API call is not supported yet", "code": 500}

—
Reply to this email directly or view it on GitHub
<#442 (comment)
.

—
Reply to this email directly or view it on GitHub
#442 (comment).

—
Reply to this email directly or view it on GitHub
#442 (comment).

[slivne]

@pdziepak I know we have talked about this - is there still something missing for this ?

[slivne]

@pdziepak , @gleb-cloudius is there still an opem item here ?

[pdziepak]

On 18 February 2016 at 00:49, slivne notifications@github.com wrote:

@pdziepak https://github.com/pdziepak , @gleb-cloudius
https://github.com/gleb-cloudius is there still an opem item here ?

Drain is implemented and working correctly. There is a regression that
causes segmentation fault when shutting down drained node, but I've just
sent a patch to fix it.