I documented this a few years ago, see the blue note here:
https://docs.ceph.com/en/latest/rados/operations/placement-groups/#viewing-pg-scaling-recommendations
> Hi guys.
>
> I get no output from:
> -> $ ceph osd pool autoscale-status
>
> I read docs and I thought I should have checked:
>
> -> $ ceph osd pool get .mgr crush_rule
> crush_rule: replicated_rule
> -> $ ceph osd pool get cephfs.MONERO.data crush_rule
> crush_rule: replicated_rule_MONERO
> -> $ ceph osd pool get cephfs.MONERO.meta crush_rule
> crush_rule: replicated_rule_MONERO
Please send `ceph osd crush rule dump`
replicated_rule most likely looks like this:
{
"rule_id": 0,
"rule_name": "replicated_rule",
"type": 1,
"steps": [
{
"op": "take",
"item": -1,
"item_name": "default"
},
{
"op": "chooseleaf_firstn",
"num": 0,
"type": "host"
},
{
"op": "emit"
}
]
},
which does not specify a CRUSH device class. Do you have both HDDs and SSDs?
replicated_rule_MONERO probably looks like this:
{
"rule_id": 3,
"rule_name": "replicated_hdd",
"type": 1,
"steps": [
{
"op": "take",
"item": -2,
"item_name": "default~hdd"
},
{
"op": "chooseleaf_firstn",
"num": 0,
"type": "host"
},
{
"op": "emit"
}
]
},
Note the different in item_name: it specifies a device class via a shadow CRUSH
tree. Which means that the rules in use are seen as spanning multiple CRUSH
roots and the autoscaler "can't even".
I put in an RFE for this last year. I've asked that the .mgr pool be created
with a device class and that a health warning be raised if this situation
detected.
Depending on your constellation of rules, the best bet may be to use the CLI to
change the CRUSH rule for the .mgr pool to replicated_rule_MONERO. Send the
requested dump and I can say for sure.
That said, I'd keep rule names lowercase, I don't *know* that mixed causes a
problem, but I like to not tempt fate.
>
> All I have are above three pools - _.mgr_ created at/by deployment(in
> containers)
>
> I must be missing something else - what that might be?
> I see no clear & obvious errors in the logs.
>
> many thanks, L.
> _______________________________________________
> ceph-users mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
_______________________________________________
ceph-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
