On Mon, May 12, 2003 at 12:06:17PM -0400, Jon LaBadie wrote:
> On Mon, May 12, 2003 at 09:22:55AM -0400, Tim Champ wrote:
> > The machine worked fine until a planned downtime where I updated the PROM,
> > and added a second CPU.  Also, I split up the disks among the two SCSI
> > busses.

Ick.  Too many things changed at once :-(


> > Now, I get no errors from any disks, and I have syslog logging on
> > debug.  The CPU seems fine, as it hasn't crashed once, or had any errors.
> 
> both cpu's in use correct (psrinfo(1))
> [somewhere else he mentions iostat]

How about vmstat and/or top?  Is the CPU usage pinned?  Is there
a lot of paging?


> > I'm stumped.  One think I've noticed is that while amanda is dumping, it
> > seems to use six "dumps" which seems strange, but I've been told that it's
> > normal.  One spawns another, which spawns 4 more.

Yeah, Solaris ufsdump does that.  Its man page (Solaris 7 anyway)
says, in the "Notes" section:
|  Process Per Reel
|     Because each reel requires a new process,  parent  processes
|     for  reels  that  are  already written hang around until the
|     entire tape is written.

<aside>
According to the man page for FreeBSD 4.8's dump, it does the
same thing (though they're honest enough to call it a bug :-) If
people aren't seeing this in FreeBSD (I haven't checked it
myself), I guess either its man page is out of date, or its
code's been made smart enough to suppress the multitape behaviour
when dumping to stdout.

FreeBSD's version of the remark shows a common ancestry with the
above-quoted Solaris version, so this would appear to be old
behaviour.
</aside>


> > Also, if I run ufsdump on the machine locally, and dump from one disk to  
> > another (on the same chain), I get 3-4MB per second.  IF I do that and
> > pipe it through gzip (as amanda does) I get 400KB per second.  But, if
> > amanda does it, I get 11KB per second.  As you might think, this means it
> > takes an eternity to dump a disk.

Stab in the dark:  Is there something else (esp. something
disk-intensive) running on that box in the middle of the night?
If your manual tests are done during working hours, they wouldn't
be contending with whatever-it-is...


> Just some thoughts that probably won't get at your problem but might
> stimulate some productive thoughts.

Likewise.

--

|  | /\
|-_|/  >   Eric Siegerman, Toronto, Ont.        erics AT telepres DOT com
|  |  /
My Wine works.  However it crashes about half the time on startup.
Apparently their simulation of windoze API is getting too accurate.  :)
        - Kyle Sallee