Chris Siebenmann
#OmniOS r151014, 64G RAM, 2x1G incoming NFS from a bunch of clients, mirrored pool on two (local) 7200 RPM SATA drives, many files written.
It's definite; an #OmniOS NFS fileserver will lock up under some sustained write loads that are faster than its disks can sustain.
-
-
- View other replies
-
@thatcks Wtf is a bunch? Lock up like stop replying to the NFS clients, or lock up like delaying longer than you waited? Expected behavior! - View other replies
-
@MirageComputing NFS servers having kernel hangs is not expected behavior unless the NFS server has known bad bugs. -
@thatcks Beating a disk subsystem that you can expect 85ish iops from with 32 clients is very likely to cause the kernel to delay long times -
@MirageComputing There is a major difference between 'delay a long time' and 'lock up'. One is okay. The other is a bug. -
@thatcks No, unqualified lock up, is what someone says who didn't wait for the system to return. Packet trace, or tell how long you waited. - View other replies
-
@MirageComputing Our setup allows me to state with 100% confidence that all IO to the disks stopped in the hang. Also: hours (test system).
-
Mirage Computing
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.