Closed
Bug 838159
Opened 12 years ago
Closed 12 years ago
Additional master server for Socorro Postgres databases
Categories
(Data & BI Services Team :: DB: MySQL, task)
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: selenamarie, Assigned: mpressman)
References
Details
The strange failure of the drive in master01 brought to light a failure mode we'd like to avoid:
* Secondary database goes offline or is unavailable
* Our master database fails and we need to fail over, or restore from a backup
We would like a third system that is the same as master01/master02 for failover.
Because we've made RO access to the secondary system available, people are building applications that connect only to it for data. We need a third system running as a replica online at all times supporting the RO zeus node.
Reporter | ||
Updated•12 years ago
|
Summary: Additional master server for Socorro databases → Additional master server for Socorro Postgres databases
Comment 1•12 years ago
|
||
Is this third system running as a replica this? https://bugzilla.mozilla.org/show_bug.cgi?id=813317
Does that count as n+1 or do we need another one?
Reporter | ||
Comment 2•12 years ago
|
||
(In reply to Sheeri Cabral [:sheeri] from comment #1)
> Is this third system running as a replica this?
> https://bugzilla.mozilla.org/show_bug.cgi?id=813317
>
> Does that count as n+1 or do we need another one?
We need another one. The Reporting replica is going to be used for a different purpose - long running experimental queries and will have a different configuration.
Comment 3•12 years ago
|
||
Awesome! Wasn't sure if it could also be used as read-only. Sounds like no.
I'm all for n+1 configuration!
cc'ing Corey for hardware ordering.
Reporter | ||
Comment 4•12 years ago
|
||
(In reply to Sheeri Cabral [:sheeri] from comment #3)
> Awesome! Wasn't sure if it could also be used as read-only. Sounds like no.
It can be used as a RO, but the query timeout will be set to 5 min. The query timeout on the reporting replica is going to be much longer - possibly an hour or more.
> I'm all for n+1 configuration!
>
> cc'ing Corey for hardware ordering.
Woot.
Comment 5•12 years ago
|
||
Blade server shipped via Fed Ex 294466306093837 ETA 2/18
Storage Blades shipped via fed Ex 812085414085796 ETA 2/15
Server hard drives (2) shipped via Fed Ex 812085414085864 ETA 2/15
Storage blade hard drives (12) shipped via Fed Ex 294466306090096 ETA 2/19
Comment 6•12 years ago
|
||
Removing Rich - the server is now racked, so it's ready for SRE kickstarting and then puppetizing. Matt - I'm leaving this in your hands.
Updated•12 years ago
|
Assignee: server-ops-database → mpressman
Assignee | ||
Comment 7•12 years ago
|
||
This is up and running as socorro3.db.phx1 and is in puppet
Assignee | ||
Updated•12 years ago
|
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Updated•10 years ago
|
Product: mozilla.org → Data & BI Services Team
You need to log in
before you can comment on or make changes to this bug.
Description
•