On 4 February 2014 08:43, Lupco Trpeski <[email protected]> wrote: > HI Guys, > > Has any one seen weird disk performance issues after updating the entire > package. > > I'm running a HP N36L Microserver booting off a Patriot 32GB USB stick. > Performance isn't amazing at the best of times but it normally runs fairly > well. Basically once you do a pkg update --accept entire and it boots into > the alternate be, wait times in wsvc_t (according to iostat -xn 1 100) goes > up to about 3000ms every third interval. > > Has anyone seen any behavior like that before? > > What is the rpool capacity before creating the updated BE and as compared with after? If it gets a lot fuller perhaps you are hitting the zfs slowdown on fuller pools (which is much improved in the most recent releases).
How many old BEs are there - can you trim the oldest? Gavin
_______________________________________________ msosug mailing list [email protected] http://mexico.purplecow.org/m/listinfo/msosug
