New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 607219 link

Starred by 1 user
Status: WontFix
Owner:
█ dshi@chromium.org
Last visit > 30 days ago
Closed: Jun 2017
Cc:
█ fdeng@chromium.org
akes...@chromium.org
█ shuqianz@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

Cros lab Master Database doesn't have automatic failover

Project Member Reported by autumn@chromium.org, Apr 27 2016 
From chat with dshi, can close if this has been done already. 

master database - has backup with the slaves. Not automatic failover but the data will be preserved. Would have to make changes like moving it to a cluster in order to do automatic failover. 

+ aviv FIY
+ Charlene in case this overlaps with her shard work

Comment 1 by dshi@chromium.org, May 2 2016

Cc: fdeng@chromium.org
+fdeng who setup the failover pair.

I think automatic failover will be a bit challenge. Couple reasons:
1. Currently the master db IP is hard coded in shadow config. 
2. For a db instance to change role from slave to master, there are some manual actions required.

Comment 2 by fdeng@chromium.org, May 2 2016 

There are a series of manual actions like Dan said to promote a slave to master which requires coordination between two slaves an a master, which are challenging to automate correctly.

I think for long term, the better direction to go would be 
  - shard all platforms
  - move master db to cloud sql (since no master scheduler is needed anymore, so we won't be that sensitive about latency)

Comment 3 by autumn@chromium.org, May 4 2016

Components: Infra>Client>ChromeOS

Comment 4 Deleted

Comment 5 by dshi@chromium.org, Jun 16 2017

Status: WontFix (was: Untriaged)
master db shall be in CloudSQL, no need to automatic failover.
►

Sign in to add a comment