Reply
Contributor
Posts: 21
Registered: ‎2009-09-11

4.1.8 T9 AFP dies overnight

AFP works for me, including with Time Machine, in T9. However, after some amount of time the AFP service dies. I'm still able to connect via CIFS, but AFP will not be available. If I restart the AFP service, it will be working again.

I'm not sure if it dies on its own, or Time Machine does something that kills it.
Contributor
Posts: 21
Registered: ‎2009-09-11

Re: 4.1.8 T9 AFP dies overnight

Looking at the logs, the timing of the death appears to coincide with the time when cron.daily runs.


Aug 16 06:08:01 ReadyNAS /USR/SBIN/CRON[15913]: (root) CMD ( /usr/bin/empty_exim &>/dev/null)
Aug 16 06:23:01 ReadyNAS /USR/SBIN/CRON[15937]: (root) CMD ( /usr/bin/empty_exim &>/dev/null)
Aug 16 06:25:01 ReadyNAS /USR/SBIN/CRON[15945]: (root) CMD (test -e /usr/sbin/anacron || nice run-parts --report /etc/cron.daily)
Aug 16 06:38:01 ReadyNAS /USR/SBIN/CRON[16129]: (root) CMD ( /usr/bin/empty_exim &>/dev/null)
Aug 16 06:53:01 ReadyNAS /USR/SBIN/CRON[16147]: (root) CMD ( /usr/bin/empty_exim &>/dev/null)



Aug 19 06:02:17.230217 afpd[8328] {auth.c:944} (N:AFPDaemon): AFP logout by ReadyNAS
Aug 19 06:02:17.277412 afpd[8328] {afp_dsi.c:99} (N:AFPDaemon): AFP statistics: 28326.31 KB read, 129318.83 KB written
Aug 19 06:02:17.277672 afpd[8328] {dircache.c:615} (I:AFPDaemon): dircache statistics: entries: 296, lookups: 7834, hits: 6631, misses: 307, added: 1192, removed: 896, expunged: 896, evicted: 0
Aug 19 06:02:17.278133 afpd[8328] {afp_dsi.c:570} (N:AFPDaemon): done
Aug 19 06:02:17.312803 afpd[25526] {main.c:216} (I:AFPDaemon): child[8328]: done
Aug 19 06:25:16.834827 afpd[25526] {main.c:148} (N:AFPDaemon): AFP Server shutting down on SIGTERM
Regular Contributor
Posts: 494
Registered: ‎2010-08-27

Re: 4.1.8 T9 AFP dies overnight

This sounds a lot like what was happening to my Pro.

I had the initial releases of the add-ons Dns-O-Matic and OpenDNS installed. After removing them and rebooting the issue went away.
Charles LaCour
clacour_AT_gmail.com
Contributor
Posts: 21
Registered: ‎2009-09-11

Re: 4.1.8 T9 AFP dies overnight

I don't have those add-ons, but I think I might have solved this issue.

It seems that if afpd has been started from the web administration interface, so that it's a child of the web server, it dies overnight when cron.daily runs. However, if I reboot, such that afpd has been started by the system and not by the web server, it doesn't do this any more.

I was seeing the behavior repeatedly because I was fixing it by restarting afpd only via the admin interface.
NETGEAR Employee
Posts: 3,207
Registered: ‎2004-11-19

Re: 4.1.8 T9 AFP dies overnight

That's right, we recently found that when netatalk starts, it doesn't fully daemonize. It does detach from the parent, but it stays in the same process group. But they are aware of it now and are in the process of fixing it.
Regular Contributor
Posts: 494
Registered: ‎2010-08-27

Re: 4.1.8 T9 AFP dies overnight

OK, that explains the issue I had as well. So it wasn't the old plugin's after all.
Charles LaCour
clacour_AT_gmail.com
Discussion Stats
  • 5 replies
  • 177 views
  • 0 kudos
  • 3 in conversation
Top Contributors