9.2-RC3 locks up on Proliant DL165

Z

ziyanm

I tested out 9.2-RC3 on a HP Proliant DL165 that has been running 9.1 and found it locks up a couple of minutes after boot. I needed to physically power cycle the machine to get it to reboot. I detected no panic found no (text)dump afterwards.

The system has an 8-core Opteron 6136, 8G 8 G(i)B RAM, 4-port Intel NIC (igb), 3x SATA in RAID-Z ZFS root pool. It runs NFS, LDAP, DNS, DHCP and a lightly loaded web server in jails on 9.1-RELEASE.

I compiled 9.2-RC from source using clang and the same kernel configuration include I'd used for 9.1 (ALTQ, IPSEC, and nodevice unneeded drivers.) Has anyone experienced similar issues? How do I report this when I can't get core dump when it happens?
 
I found a similar situation on an 8-core 16-thread Dell. Zero debug information available. Just locked solid, and I have to power cycle it.

This machine hasn't been running 9.1, but similar hardware configurations are running fine. This is a test bed, but at this point I am thinking something is up with 9.2. I can throw some more machines at 9.2 to prove this over the weekend.
 
Thanks for the feedback everyone. @grabes, can you post your hardware and system configuration to see what we have in common?

I have a bit more puzzling information, which indicates it *may not* be a kernel issue. I use two separate datasets tank/91RELEASE and tank/92RELEASE for testing. I copied over the 9.2 kernel to /boot/92kernel on tank/91RELEASE, which is the pool's bootfs. Now I break into the loader prompt and do the following
Code: 
unload
set kernel=92kernel
set vfs.root.mountfrom=zfs:tank/92RELEASE
boot
Booting this way has 9.2-RC3 running fine for a couple of hours now. However, If I do zpool set bootfs=tank/92RELEASE tank and reboot with the very same kernel it locks up solid a couple of minutes after boot!
 
Last edited by a moderator:
You must log in or register to reply here.
Back
Top