On 01/05/2017 at 02:02 PM, Dave Young wrote:
On 01/05/17 at 10:51am, Xunlei Pang wrote:
> On 01/03/2017 at 02:41 PM, Baoquan He wrote:
>> Hi Xunlei,
>>
>> As we discussed in meeting, I personally tend to agree not do it this
>> way. Mainly because it's a corner case and there isn't a clear judgement
>> to figure out what system could meet the corner case. So it will spread
>> out into each system, and may bring later maintenance cost though
>> unpredicted now. Maybe can fix it later when customer has a specific
>> request to fix this.
>>
>> Warn user of the danger, or let customer adjust nr_cpus manually, either
>> is fine to me.
> Hi Baoquan,
>
> Thanks for the comments.
> Yes, I'd prefer throwing some warnings when service restarts to tip users.
Xunlei, since we have detected that kdump kernel can not boot with
current values, we still use it with a warning, one may not notice it.
It will 100% fail to boot, so maybe error out is better.
It's an estimated figure, it's very likely that it will fail but is not 100%
sure.
In addition, for multiple cpu systems, for interrupts bound to some cpu mask,
one irq actually consumes multiple vectors, it allocates one vector on each cpu
contained in the mask.
If it happens to boot with current value, then error out would be a false-positive
decision.
Regards,
Xunlei