Hello,

        My server runs a 11.0_BETA system. Last night, server has crashed
(power outage during a storm). I have restarted this morning this server
with a recent kernel:

legendre# uname -a
NetBSD legendre.systella.fr 11.0_BETA NetBSD 11.0_BETA (CUSTOM) #4: Sun
Oct 19 20:12:55 CEST 2025
[email protected]:/usr/src/netbsd-11/obj/sys/arch/amd64/compile/CUSTOM
amd64

        For a long time, I use readprio queues for disks. After reboot, all
diskless systems ran slowly and load average on server was very high
(greater than 50).

        I have checked strategy and all volumes return fcfs.

        Of course, readprio was built into kernel:

legendre# sysctl kern.bufq.strategies
kern.bufq.strategies = readprio priocscan fcfs disksort

        I have manually change disk strategies to readprio and, now, server
runs fine again.

        Since I use NetBSD, I believed that disk strategies were persistent.
Should I add these parameters in /etc/rc.local ?

        Regards,

        JB

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to