In message <[EMAIL PROTECTED]>, Matt Dillon writes:
>:I'm seeing this kind of hang about twice a day on my build-box.
>:Any clues ? insights ?
> When did this start occuring?
Can't say for sure, it's certainly noticeable today, but I have not
had much FreeBSD time recently so I would be hard pressed to say
more definitively when it started.
> But -current has generally been very unstable for me
> for months now and that makes it extremely difficult to isolate problems.
Yeah, I'm increasingly in favour of a "bake-out" period where no new
subsystems are SMPng'ed until -current is stabilized. Sos tells me that
running SMP is an instant death request these days. My SMP systems are
not even powered on right now.
> What's the filesystem configuration being hit by the compile? Are you
> running any MFS or VN paritions? Is softupdates turned on on the
> underlying filesystem?
Two ATA disks one CCD striped, softupdates.
> On the face of it it looks like a bawrite() is being turned into a
> bdwrite() somehow and that spec_fsync() is hitting its 'goto loop'
> restart code and looping on the same buffer... which can happen if
> the bawrite() is being turned into a bdwrite(). It looks like
> spec_fsync() is in an infinite loop.
There is no other signs of trouble, processes hang when they attempt
some (but maybe not all) kind of I/O.
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
[EMAIL PROTECTED] | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message