Skip to Content.
Sympa Menu

en - Re: [en@sympa] sympa behaviour when database is down

Subject: The mailing list for listmasters using Sympa

List archive

Chronological Thread  
  • From: Darrin de Groot <address@concealed>
  • To: Phil Stracchino <address@concealed>, address@concealed
  • Subject: Re: [en@sympa] sympa behaviour when database is down
  • Date: Tue, 11 Oct 2022 08:22:37 +0200

On 06/10/2022 19:15, Phil Stracchino wrote:
On 10/6/22 03:32, Darrin de Groot wrote:
Hi Soji,

Thanks very much for your reply - it helped me to go back and look again
at the logs, as our mariadb/mysql error is slightly different. We are
running a galera cluster and it seems that the service was, in fact, up,
but that the cluster was not ready to answer queries - resulting in an
"Unable to execute SQL statement", rather than simply not being able to
connect:


Darrin,
Are you using ProxySQL or HAProxy in front of your cluster? HAProxy combined with a clustercheck script of one variety or another SHOULD enable you to maintain database connectivity to a Galera cluster as long as *any one node* is available to serve queries.  (Also remember that Galera can be configured to serve *read-only* queries from a node *even while serving an SST* and not available for writes.)



Hi Phil,

We are using HAProxy and have a clustercheck script - the database connectivity seems to have been maintained. We may simply need to, as you suggest, look at the galera config to serve read-only while serving and SST. Very good point - thank you, I'll check our config.

Regards,

Darrin.




Archive powered by MHonArc 2.6.19+.

Top of Page