tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Reclaiming vnodes

On Sep,Tuesday 15 2009, at 1:43 AM, Wolfgang Solfrank wrote:


Vnodes are only allocated and there is no vnode recycling. If number of used vnodes in a system is > than kern.vnodes_num_hiwat (in percents of maxvnodes) vrele thread will be woken up and it will start with releasing of free vnodes until number of used vnodes is < than kern.vnodes_num_lowat.

And how will that help the situation?

As far as I understand the situation the system is in a state that it can't handle the amount of resource requests forwarded to it, and thus will probably need to sleep on the request for some new vnode anyway. Maybe you are shifting the problem a bit further along, but it doesn't help in principle anyway.

Yes it helps because vnodes are not released in zfs context but in a different one where is only VOP_RECLAIM called without any other zfs locks held. We shift vnode reclaiming from fs context to vrele thread.



Home | Main Index | Thread Index | Old Index