For various reasons, I'm using illumos from openindiana under
virtualbox. As a result of doing so, I'm seeing some very strange
behaviour.
For example, the system is currently doing nothing BUT top tells
me that the system is spending about 10% of tis time in the kernel.
"vmstat 1" similarly reports the system time as up to 20% when it
is doing nothing other than running vmstat.
Anyone want to volunteer a dtrace script to analyse this further?
Along with that, when booting the system, I see microcode update
failure messages:
Feb 4 16:41:42 oi151a unix: [ID 631382 kern.warning] WARNING: cpu1:
failed to update microcode from version 0x0 to 0x11
... which I take to mean that the virtual CPUs aren't updateable.
The rest of the CPU probing goes fine:
Feb 4 16:41:42 oi151a unix: [ID 950921 kern.info] cpu1: x86 (chipid 0x0
GenuineIntel 106A5 family 6 model 26 step 5 clock 2633 MHz)
Feb 4 16:41:42 oi151a unix: [ID 950921 kern.info] cpu1: Intel(r)
Core(tm) i7 CPU 920 @ 2.67GHz
Feb 4 16:41:42 oi151a unix: [ID 557947 kern.info] cpu1 initialization
complete - online
Feb 4 16:41:40 oi151a ahci: [ID 296163 kern.warning] WARNING: ahci0:
ahci port 0 has task file error
Feb 4 16:41:40 oi151a ahci: [ID 687168 kern.warning] WARNING: ahci0:
ahci port 0 is trying to do error recovery
Feb 4 16:41:40 oi151a ahci: [ID 693748 kern.warning] WARNING: ahci0:
ahci port 0 task_file_status = 0x441
Feb 4 16:41:40 oi151a ahci: [ID 332577 kern.warning] WARNING: ahci0:
the below command (s) on port 0 are aborted
Feb 4 16:41:40 oi151a ahci: [ID 117845 kern.warning] WARNING: satapkt
0xffffff01482538c8: cmd_reg = 0xe3 features_reg = 0x0 sec_count_msb =
0x0 lba_low_msb = 0x0 lba_mid_msb = 0x0 lba_high_msb = 0x0 sec_count_lsb
= 0x0 lba_low_lsb = 0x0 lba_mid_lsb = 0x0 lba_high_lsb = 0x0 device_reg
= 0x0 addr_type = 0x0 cmd_flags = 0x4010011
Feb 4 16:41:40 oi151a ahci: [ID 657156 kern.warning] WARNING: ahci0:
error recovery for port 0 succeed
But perhaps the strangest behaviour that I've seen is on a
couple of occasions when I've woke Windows up from an overnight
sleep without doing a 'pause' in VirtualBox. The system went
to sleep whilst I was doing a build. At first the response was
less than immediate in putty but ok on the console, so I closed
putty and tried again. After that, both using putty and the
console was abysmal however the build continued to run in the
background, with the disk activity light flashing on virtualbox.
As an example of this, I was running "vmstat 1" on the console
and after printing the first line, it was rather sporadic in
printing successive lines.
Unfortunately even though I forced a system panic and crash dump,
I'd forgotten to create /var/crash & friends and thus lost the
system. Has any thought been given to forcing the creation of the
directories required to store crash dumps from a default install?
Finally, whilst running "iostat 1", there are some rather odd values
that turn up in some of the disk activity fields, e.g.:
tty sd0 cpu
tin tout kps tps serv us sy wt id
0 118 0 0 0 0 5 0 95
0 190 0 0 0 0 6 0 94
0 119 0 0 0 0 6 0 94
0 118 0 0 0 0 6 0 94
0 118 0 0 0 0 6 0 94
0 119 0 0 0 0 9 0 91
0 117 0 0 0 0 8 0 92
0 118 0 0 0 0 5 0 95
0 119 0 0 0 0 5 0 95
0 119 0 0 0 0 1 0 99
0 117 0 0 0 0 3 0 97
0 117 69 7 2635249153383 0 5 0 95
0 127 0 0 0 0 6 0 94
0 119 0 0 0 0 11 0 89
0 118 0 0 0 0 6 0 94
0 117 0 0 0 0 10 0 90
0 119 248 72 256204778765 0 26 0 74
0 125 0 0 0 0 4 0 95
Darren
-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175430-2e6923be
Modify Your Subscription:
https://www.listbox.com/member/?member_id=21175430&id_secret=21175430-6a77cda4
Powered by Listbox: http://www.listbox.com
