On 2010-05-05 12:18, Florian Haas wrote:
> On 2010-05-05 12:03, Dag Stenstad wrote:
>>> that's a good idea, though I'm not sure if I like the implementation.
>>> Rather than specify "parent_domain", why not introduce a boolean
>>> parameter which if set, causes the migration_uri to be constructed from
>>> a host(1) lookup of $OCF_RESKEY_CRM_meta_migrate_target, so you retrieve
>>> the FQDN from that?
>>> Or maybe it's a good idea to always do that?
>> Yes, I agree that the implementation might not be the best. It seems
>> like you might mean doing something like this?
>>
>> target_node="$OCF_RESKEY_CRM_meta_migrate_target"
>> if ocf_is_true $OCF_RESKEY_migrate_use_fqdn; then
>>     target_node=$(host $OCF_RESKEY_CRM_meta_migrate_target | \
>>     awk '{ print $1; }')
>> fi
>>
>> Alternatively, one could also do something like this to always use the
>> local domain:
>>
>> target_node="$OCF_RESKEY_CRM_meta_migrate_target"
>> if ocf_is_true $OCF_RESKEY_migrate_use_domain; then
>>     target_node=${OCF_RESKEY_CRM_meta_migrate_target}.$(hostname -d)
>> fi
>>
>> To be honest, I'm not sure wich one I prefer. What I want to accomplish
>> is to have a FQDN for migration, so that my SSL certificates match..
>>
>> Opinions? Both solves my problem.
> 
> I'd prefer the first option. You might have a cluster consisting of
> alice.foo and bob.bar. I concede that's far-fetched, but option #1 takes
> care of that while #2 does not, and neither seems to be causing much pain.

Actually, on second thought #2 does cause pain, as on some systems
"uname -n" does contain the domain name, meaning
$OCF_RESKEY_CRM_meta_migrate_target will too, and if you just blindly
append the domain name you'll duplicate it.

Cheers,
Florian


Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/

Reply via email to