Re: System reboots building mysql51-client with ZFS v28

On Thu, Sep 9, 2010 at 4:04 PM, Xin LI <delphij_at_delphij.net> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
>
> On 2010/09/09 15:33, Xin LI wrote:
>> On 2010/09/09 15:28, Rob Farmer wrote:
>>> On Thu, Sep 9, 2010 at 15:14, Xin LI <delphij_at_delphij.net> wrote:
>>>> -----BEGIN PGP SIGNED MESSAGE-----
>>>> Hash: SHA256
>>>>
>>>> Hi, Rob,
>>>>
>>>> On 2010/09/09 15:01, Rob Farmer wrote:
>>>> [...]
>>>>>> Just a guess - are you compiling under X?  If so what about if you
>>>>>> compile under text mode?
>>>>>
>>>>> Nope - this is a new install and X isn't running yet. It is installed,
>>>>> but not started yet. I ran across this building dependencies for KDE
>>>>> (why a desktop requires MySQL server is beyond me, but anyways).
>>>>
>>>> I just tried to compile MySQL 5.1 client but it passed on my test system
>>>> (the compile time option is slightly different as the VFS debugging is
>>>> not yet enabled).  I'll do another round of test with these VFS options
>>>> enabled.
>>
>>> Ok - so I shut the system down for about 15 minutes and then tried
>>> again and it worked.
>>
>>> So maybe it was heat or something - I had reproduced it about 5 times
>>> back to back at the exact same spot. I didn't think this laptop had
>>> cooling problems but I will keep a closer eye on it in the future.
>>
>> Oh...  If it's so reproducible like this way then I'm more inclined to
>> believe that it's a software issue.  I'll try to see if I can reproduce
>> the problem myself anyways, thanks for letting us know :)
>
> Ok mine went through with these recommended debugging options.
>
> Note that on my kernel I have bumped FULLGRAPH_SBUF_SIZE to 131072 in
> /sys/kern/subr_witness.c.  I'm not sure if it's related but without that
> I used get a few panics in the past.

As of r212370 on CURRENT this is no longer necessary, as the
sbuf_printf() will drain to the sysctl struct.  The define and its use
are gone.

Thanks,
matthew