Will Berkeley has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/11486 )

Change subject: [docs] Add tip on dealing with planned TS downtime
......................................................................


Patch Set 1:

(4 comments)

http://gerrit.cloudera.org:8080/#/c/11486/1/docs/administration.adoc
File docs/administration.adoc:

http://gerrit.cloudera.org:8080/#/c/11486/1/docs/administration.adoc@1126
PS1, Line 1126: If a single tablet server is brought down temporarily in a 
healthy cluster, all
> Is the first sentence relevant? Maybe just talking about the replication as
It establishes the consequences of downtime for a single tablet server, in 
full, rather than only highlighting the downside. The point being that a user 
investigating what will happen to their Kudu cluster when they bring down one 
tablet server can read this section and understand the implications of their 
plan.

I also wanted to be careful to establish the idea that the downtime is planned 
and done in a healthy cluster-- I do not want users to think they can apply 
this advice if they have unexpected downtime.


http://gerrit.cloudera.org:8080/#/c/11486/1/docs/administration.adoc@1140
PS1, Line 1140:  on each tablet server:
> "for each tablet server", since we don't have to be local to run the comman
Done


http://gerrit.cloudera.org:8080/#/c/11486/1/docs/administration.adoc@1151
PS1, Line 1151: $ sudo -u kudu kudu tserver set_flag <tserver_address> 
follower_unavailable_considered_failed_sec <original_value>
> Do you think it's worth adding a sentence to call out the `get_flags` tool
I don't want to, for a couple of reasons:

0. It's safe to assume pretty much everyone uses the default.
1. Technically the flag can differ between tablet servers. If I include it in 
the instructions and we insist on prevision it's N more commands people need to 
run when they probably don't...they just need the default of 300 N times.
2. The command to get the value is a little gross

    sudo -u kudu kudu tserver get_flags <tserver_addr> --all_flags --format=csv 
| grep follower_unavailable_considered_failed_sec | cut -d',' -f2


http://gerrit.cloudera.org:8080/#/c/11486/1/docs/administration.adoc@1157
PS1, Line 1157: On Kudu versions prior to 1.8, the `--force` flag must be 
provided in the above
> maybe prepend "NOTE:"
Done



--
To view, visit http://gerrit.cloudera.org:8080/11486
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I55a992a00f35945187e02c55594edc6e261a72c4
Gerrit-Change-Number: 11486
Gerrit-PatchSet: 1
Gerrit-Owner: Will Berkeley <[email protected]>
Gerrit-Reviewer: Andrew Wong <[email protected]>
Gerrit-Reviewer: Grant Henke <[email protected]>
Gerrit-Reviewer: Will Berkeley <[email protected]>
Gerrit-Comment-Date: Thu, 20 Sep 2018 21:18:00 +0000
Gerrit-HasComments: Yes

Reply via email to