Mailing List Archive

Re: [Linux-HA] Antw: Re: Q: What type of dependency is "colocation"?
On 2011-07-11 17:08, Ulrich Windl wrote:
>>>> Florian Haas <florian.haas@linbit.com> schrieb am 11.07.2011 um 16:46 in
> Nachricht <4E1B0CD4.80003@linbit.com>:
>> On 2011-07-11 15:14, Ulrich Windl wrote:
> [...]
>>> OK:
>>> primitive prm_rksapr00_ping ocf:pacemaker:ping \
>>> params ... \
>>
>> Any specific reason why you're cutting configuration parameters out?
>
> They are not relevant and only add water to the soup.

But they're... never mind, you'd probably interpret it as another case
of not liking your opinion.

>>> op monitor interval="300s" timeout="60" \
>>> op start interval="0" timeout="60" \
>>> utilization utl_cpu="1" utl_ram="1" \
>>> meta priority="2050" target-role="Started"
>>> group grp_rksapr00 prm_rksapr00_ip_1 ... \
>>
>> Here too?
>
> It is not important how many members the group has, right?
>
>>
>>> meta priority="2000" resource-stickiness="100000"
>> target-role="Started"
>>> order ord_rksapr00_ping_after_saprouter inf: grp_rksapr00 prm_rksapr00_ping
>>> colocation col_rksapr00_saprouter_ping inf: grp_rksapr00 prm_rksapr00_ping
>>
>> Why is the ping resource not cloned, and what is this colocation
>> supposed to achieve?
>
> The ping should actually just ping to keep a network connection alive. The other is just a side effect. The ping should use the source IP address (not possible with standard Linux ping)

What exactly isn't possible with "standard Linux ping"? What about "ping
-I"?

> that prm_rksapr00_ip_1 provides. Originally I wanted to write an RA for it, bit the pingd seems to do what I need.

Does "crm ra info ocf:heartbeat:IPsrcaddr" not help at all?

>>> OK, so it seems you implemented a strange kind of c "colocation" that is
>> part of co-location, and part of "depends_on". I'd like to have clean and
>> separate implementations:
>>
>> Excellent. Send a patch!
>
> It's like Christmas wishes: we all have to wait until Christmas. ;-)

No, you'll have to wait until someone implements the functionality you
want. Or you implement it yourself.

>>> For co-location: If "A is near B", obviously "B is near A", so co-location is
>> symmetric by nature.
>>>
>>> For "depends_on": if "A depends_on B" it makes not much sense if "B
>> depends_on A" (is this antisymmetric?)
>>>
>>> Your implementation mixed both, a symmetric and a non-symmetric relation.
>> Naturally this causes problems.
>>
>> Out of curiosity, to whom does the possessive pronoun apply?
>
> Plural meaning: Those whose names I don't know, but are responsible for the implementation.

Ah.

>>> Well actually I feel you just don't like any opinion others than your own.
>>
>> Says exactly who, about whom?
>
> OK, singular this time: The author of the message.

It takes one to know one, doesn't it?

Florian
Re: [Linux-HA] Antw: Re: Q: What type of dependency is "colocation"? [ In reply to ]
Apologies to ha-wg-technical subscribers, posted my reply to the wrong list.

Cheers,
Florian
Re: [Linux-HA] Antw: Re: Q: What type of dependency is "colocation"? [ In reply to ]
>>> Florian Haas <florian.haas@linbit.com> schrieb am 11.07.2011 um 17:16 in
Nachricht <4E1B13C5.4090505@linbit.com>:
> On 2011-07-11 17:08, Ulrich Windl wrote:
[...]
> What exactly isn't possible with "standard Linux ping"? What about "ping
> -I"?

Hi!

Option "-I" in Linux (LES11 SP1, iputils-ss021109-292.28.1) ping allows to specify the interface, not the IP address that is used for outgoing ping packets. If you added multiple IP adresses to one interface (via IPaddr2), it's not obvious which address is used. When I tried to use the "interface label" (like "bond0:a12"), the part after the colon was effectively ignored (verified by specifiying non-existing labels).

Regards,
Ulrich


_______________________________________________
ha-wg-technical mailing list
ha-wg-technical@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/ha-wg-technical