TOPIC:

Hi,

I've just configured my HA-Lizard Pool to send email alerts.

I've just noticed that its sending me this alert:

"write_pool_state: Error retrieving autopromote_uuid from pool configuration"

Anyone know what this is / what I need to do to fix it?

Regards,

Adam

autopromote_uuid holds the UUID of a slave that is selected to take over as pool master in the event that the master host fails.

This alert may be sent once during an HA event where the slave promotes itself to a master and no other pool members are present.

It could also occur hourly if the pool has only a master and no slave hosts present.

We should probably be suppressing this alert in either of the situations above.

Regardless, if you pool has at least 2 active members (1 master and 1 slave) then you should not be seeing this message. If you are, then something could be wrong with your configuration or pool.

sc wrote: autopromote_uuid holds the UUID of a slave that is selected to take over as pool master in the event that the master host fails.

This alert may be sent once during an HA event where the slave promotes itself to a master and no other pool members are present.

It could also occur hourly if the pool has only a master and no slave hosts present.

We should probably be suppressing this alert in either of the situations above.

Regardless, if you pool has at least 2 active members (1 master and 1 slave) then you should not be seeing this message. If you are, then something could be wrong with your configuration or pool.

That's worrying then - I have a 2 node Pool (1 master and 1 slave). I am receiving these emails every hour.

Any ideas what could be wrong with my configuration / pool? Or where to look first for diagnosis?

Try issuing this command from the shell. You should get the UUID of the slave back.

xe pool-param-get uuid=$(xe pool-list --minimal) param-name=other-config param-key=autopromote_uuid

Post the results if you see an error. It could be that the database field used to store the value is missing from your configuration. If so, try this from the shell to add it.

xe pool-param-add uuid=$(xe pool-list --minimal) param-name=other-config autopromote_uuid=""


if you've confirmed that the autopromote_uuid DB field is present, then post about 1 minute of log from the log file of the master and the slave to further review.

Hi,

I ran the commands - the first one returned just a blank line (see below for captured output). So I then ran the command to add the autopromote_uuid and it returned that the key is already there (see below)

[root@MY-POOL-MASTER ~]# xe pool-param-get uuid=$(xe pool-list --minimal) param-name=other-config param-key=autopromote_uuid

[root@MY-POOL-MASTER ~]# xe pool-param-add uuid=$(xe pool-list --minimal) param-name=other-config autopromote_uuid=""
You tried to add a key-value pair to a map, but that key is already there.
type: pool
param_name: other_config
uuid: 243e104a-26fe-7d10-49fb-15ebf95b6e2f
key: autopromote_uuid
[root@MY-POOL-MASTER ~]#

which logs do you need for further analysis?

thanks for your help,

Adam

type "ha-cfg log" on both the master and the slave. This will output real time log to stdout. Capture about 1 minute from the master and the slave and attach to this thread.

thanks