This is likely a linux or ssh problem, but perhaps someone else on this
list may have encountered it. I have run about 15,000 instances of
rsync in the last year, (about 500 total hours of rsyncing) using ssh
as a transport mechanism ( actually, I am running the dirvish backup
wrapper, www.dirvish.org, around rsync). In all cases, I am pulling
filesystems from a linux client to a linux backup server. Rsync has
worked flawlessly every time. Except ...
Sometimes in my logs I see, at random, the warning "PRNG is not
seeded",
which some folks encounter when they improperly configure ssh or
/dev/urandom. This appears to emit from ssh to stderr; there is
nothing in any of the files in /var/log, or in the -vv log from rsync.
This has happened 6 times in the 12 months, twice in the last week.
Pretty rare.
Rsync is still doing its job, and completes the task on time and without
apparent data corruption. However, I am predisposed to fret about small
things, and wonder if the message indicates something needing attention.
I am willing to add more debugging flags or recompile code to track
this down, but an hour's worth of rsyncing a night is not a good
candidate for verbose tools like strace. So I'm wondering if others
have seen this, or if anyone has a well-thought-out solution for
locating the precise cause of the "PRNG is not seeded" message.
Keith
--
Keith Lofstrom keithl@keithl.com Voice (503)-520-1993
KLIC --- Keith Lofstrom Integrated Circuits --- "Your Ideas in
Silicon"
Design Contracting in Bipolar and CMOS - Analog, Digital, and Scan ICs