RE: Bacula: VERY SLOW on SAME host

Thanks to Pawel Jakub Dawidek, I found the empty "arp -an" output
bug. The bug was introduced in my last commit that attempted at 
providing some level of binary compatibility.

Please sync file ./src/sys/net/rtsock.c to svn r187328 

Thanks,

-- Qing


-----Original Message-----
From: Larry Rosenman [mailto:ler_at_lerctr.org]
Sent: Thu 1/15/2009 8:07 PM
To: Li, Qing
Cc: freebsd-current_at_freebsd.org; freebsd-net_at_freebsd.org
Subject: RE: Bacula: VERY SLOW on SAME host
 

On Thu, January 15, 2009 6:46 pm, Li, Qing wrote:
> This is a known issue and it's easily reproducible using the
> netperf tool.
>
> I am working on a permanent solution.
>
> In the meantime you can use the following workaround, as suggested
> by Kip:
>
> 	route add -host (if-ip) -iface lo0
>

Thanks, Qing!

If you have code you'd like me to test, feel free.

On a related note, is the arp table supposed to be empty now?

$ arp -an
$ ping 192.168.200.5
PING 192.168.200.5 (192.168.200.5): 56 data bytes
64 bytes from 192.168.200.5: icmp_seq=0 ttl=255 time=0.810 ms
64 bytes from 192.168.200.5: icmp_seq=1 ttl=255 time=0.292 ms
^C
--- 192.168.200.5 ping statistics ---
2 packets transmitted, 2 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 0.292/0.551/0.810/0.259 ms
$ arp -an
$ uname -a
FreeBSD borg.lerctr.org 8.0-CURRENT FreeBSD 8.0-CURRENT #1: Wed Jan 14
17:16:03 CST 2009     root_at_borg.lerctr.org:/usr/obj/usr/src/sys/BORG 
amd64
$

> -- Qing
>
>
>> -----Original Message-----
>> From: owner-freebsd-current_at_freebsd.org [mailto:owner-freebsd-
>> current_at_freebsd.org] On Behalf Of Kip Macy
>> Sent: Wednesday, January 14, 2009 8:47 PM
>> To: Larry Rosenman
>> Cc: freebsd-current_at_freebsd.org
>> Subject: Re: Bacula: VERY SLOW on SAME host
>>
>> arpv2 - add an localhost interface route as a workaround
>>
>> -Kip
>>
>> On Wed, Jan 14, 2009 at 8:35 PM, Larry Rosenman <ler_at_lerctr.org>
> wrote:
>> > Greetings,
>> >    Somethings changed (Around the 1st of the year?) where my bacula
>> jobs
>> > take FOREVER to backup on the same host
>> > with the StorageDaemon and Director.  If I change it to use
> 127.0.0.1
>> > instead of the address on the em card, it's fine.
>> >
>> > If I use the address on the em nics, it's like 4kb/sec as opposed to
>> > multi-megabytes/sec.
>> >
>> > I'm looking for how to find what broke.
>> >
>> > Other networking to the host is fine, but this is TCP within the
> same
>> host,
>> > but using the IP address on the em interface.
>> >
>> > I see the TCP Send-Q fill up, and it's extremely slow.
>> >
>> > What data do you need to help debug this?
>> >
>> >
>> >
>> > --
>> > Larry Rosenman                     http://www.lerctr.org/~ler
>> > Phone: +1 512-248-2683                 E-Mail: ler_at_lerctr.org
>> > US Mail: 430 Valona Loop, Round Rock, TX 78681-3893
>> > _______________________________________________
>> > freebsd-current_at_freebsd.org mailing list
>> > http://lists.freebsd.org/mailman/listinfo/freebsd-current
>> > To unsubscribe, send any mail to "freebsd-current-
>> unsubscribe_at_freebsd.org"
>> >
>> _______________________________________________
>> freebsd-current_at_freebsd.org mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-current
>> To unsubscribe, send any mail to "freebsd-current-
>> unsubscribe_at_freebsd.org"
>


-- 
Larry Rosenman                     http://www.lerctr.org/~ler
Phone: +1 512-248-2683                 E-Mail: ler_at_lerctr.org
US Mail: 430 Valona Loop, Round Rock, TX 78681-3893