essd wrote:
I think I've identified the cause of this issue, at least for me!
<...>
It took a couple of hours of testing, but I narrowed the problem down to the iTunes Streaming Server (Firefly). While this service is running, the system will not spin down for more than a few minutes.
I take that back. Although Firefly was definitely preventing spindown for me, there is something else preventing it too. Even with Firefly shut off and the
latest and greatest Spindown update that was posted earlier this week, I'm still seeing spin-ups for no apparent reason.
There is no file activity happening on the local network, but something is waking up the NAS. My syslog looks like this for an example period:
Oct 9 15:42:40 nas noflushd[835]: Spinning down /dev/hdc.
Oct 9 15:42:42 nas noflushd[835]: Spinning down /dev/hde.
Oct 9 15:42:52 nas noflushd[835]: Spinning up /dev/hdc after 0 minutes.
Oct 9 15:42:52 nas noflushd[835]: Spinning up /dev/hde after 0 minutes.
Oct 9 15:53:04 nas noflushd[835]: Spinning down /dev/hdc.
Oct 9 15:53:06 nas noflushd[835]: Spinning down /dev/hde.
Oct 9 15:54:37 nas noflushd[835]: Spinning up /dev/hdc after 1 minutes.
Oct 9 15:54:46 nas noflushd[835]: Spinning up /dev/hde after 1 minutes.
Oct 9 16:00:02 nas ntpdate[2024]: step time server 12.151.34.177 offset -0.130965 sec
Oct 9 16:10:08 nas noflushd[835]: Spinning down /dev/hdc.
Oct 9 16:10:10 nas noflushd[835]: Spinning down /dev/hde.
Oct 9 16:23:13 nas noflushd[835]: Spinning up /dev/hdc after 13 minutes.
Oct 9 16:23:22 nas noflushd[835]: Spinning up /dev/hde after 13 minutes.
Oct 9 16:33:19 nas noflushd[835]: Spinning down /dev/hdc.
Oct 9 16:33:21 nas noflushd[835]: Spinning down /dev/hde.
I was running tcpdump to capture traffic over this period, and I saw no traffic that *should* be waking up the NAS. Packets were of the random variety that one would expect for a network with no local activity: broadcast NetBIOS nameserver packets; ARP; broadcast IGMP; broadcast Bonjour announcements; and the one NTP request above.
I do see a correlation between ARP request/responses for the NAS and spin-ups. For example, from the above log (with MAC address sanitized for privacy):
15:42:40.871093 arp who-has 10.1.1.2 tell 10.1.1.5
15:42:40.871193 arp reply 10.1.1.2 is-at 00:0d:a2:xx:xx:xx (oui Unknown)
16:23:13.829446 arp who-has 10.1.1.2 tell 10.1.1.5
16:23:13.829547 arp reply 10.1.1.2 is-at 00:0d:a2:xx:xx:xx (oui Unknown)
10.1.1.2 is the NAS and .5 is a workstation on the network. These were the only times that I saw ARP requests for .2, so the timing seems to be more than coincidental.
Is it possible that the NAS is spinning up solely to answer ARP queries? If so, is there anything I can do to stay spun down? I'd be happy to provide the tcpdump data to the Jedis if it would help.
Thanks!
-essd