AF:
NF:0
PS:10
SRH:1
SFN:
DSR:
MID:<20081114175406.7b16925c@ripper.onstor.net>
CFG:
PT:0
S:andy.sharp@onstor.com
RQ:
SSV:exch1.onstor.net
NSV:
SSH:
R:<maxim.kozlovsky@onstor.com>
MAID:1
X-Sylpheed-Privacy-System:
X-Sylpheed-Sign:0
SCF:#mh/Mailbox/sent
RMID:#imap/andys@onstor.net@exch1.onstor.net/INBOX	0	2779531E7C760D4491C96305019FEEB5175D75748C@exch1.onstor.net
X-Sylpheed-End-Special-Headers: 1
Date: Fri, 14 Nov 2008 17:54:26 -0800
From: Andrew Sharp <andy.sharp@onstor.com>
To: Maxim Kozlovsky <maxim.kozlovsky@onstor.com>
Subject: Re: please review 31135
Message-ID: <20081114175426.10957ff6@ripper.onstor.net>
In-Reply-To: <2779531E7C760D4491C96305019FEEB5175D75748C@exch1.onstor.net>
References: <20081110174445.141b4337@ripper.onstor.net>
	<2779531E7C760D4491C96305019FEEB5175D4029BC@exch1.onstor.net>
	<20081111152935.7fdfef1e@ripper.onstor.net>
	<2779531E7C760D4491C96305019FEEB5175D402E3F@exch1.onstor.net>
	<20081114171701.09cb4974@ripper.onstor.net>
	<2779531E7C760D4491C96305019FEEB5175D757486@exch1.onstor.net>
	<20081114172321.6101b8de@ripper.onstor.net>
	<2779531E7C760D4491C96305019FEEB5175D75748B@exch1.onstor.net>
	<20081114173059.15b96493@ripper.onstor.net>
	<2779531E7C760D4491C96305019FEEB5175D75748C@exch1.onstor.net>
Organization: Onstor
X-Mailer: Sylpheed-Claws 2.6.0 (GTK+ 2.8.20; x86_64-pc-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

Hmm, well you bring up an interesting point.  Is the hosts file for the
management vsvr stored in the cluster database?  Because there is this
thing called /etc/hosts.user that is supposed to contain some
user-added hosts file entries, but for what exactly, I don't know.  And
where is the code that would search that file?

On Fri, 14 Nov 2008 17:31:52 -0800 Maxim Kozlovsky
<maxim.kozlovsky@onstor.com> wrote:

> Well for the virtual servers the hosts file is stored in the cluster
> database so it has to call the cluster database eventually.
> 
> >-----Original Message-----
> >From: Andy Sharp
> >Sent: Friday, November 14, 2008 5:31 PM
> >To: Maxim Kozlovsky
> >Subject: Re: please review 31135
> >
> >On Fri, 14 Nov 2008 17:28:16 -0800 Maxim Kozlovsky
> ><maxim.kozlovsky@onstor.com> wrote:
> >
> >> What do you do about the host names in the local host file?
> >
> >Beats me.  I assume that res_search deals with that via the
> >nsswitch.conf file.
> >
> >nfxdns_res_search() calls res_search(3) and res_init(3) which are in
> >libresolv.  Do those invoke RMC stuff?
> >
> >> >-----Original Message-----
> >> >From: Andy Sharp
> >> >Sent: Friday, November 14, 2008 5:23 PM
> >> >To: Maxim Kozlovsky
> >> >Subject: Re: please review 31135
> >> >
> >> >nfxdns_res_search().  But it doesn't seem to use RMC.  It makes a
> >> >call to SendAgileMsg() which Chris told me doesn't use RMC.  Am I
> >> >wrong?  I hope not because it's been working rather well compared
> >> >to past tries, at least up to now.
> >> >
> >> > On Fri, 14 Nov 2008 17:18:42 -0800 Maxim Kozlovsky
> >> ><maxim.kozlovsky@onstor.com> wrote:
> >> >
> >> >> Don't you use the resolver library?
> >> >>
> >> >> >-----Original Message-----
> >> >> >From: Andy Sharp
> >> >> >Sent: Friday, November 14, 2008 5:17 PM
> >> >> >To: Maxim Kozlovsky
> >> >> >Subject: Re: please review 31135
> >> >> >
> >> >> >Nope, there's another one:
> >> >> >
> >> >> >I significantly revamped the design when I was able to trace
> >> >> >through the code and determine that only the
> >> >> >libcluster_getVsId() call was doing any RMC, so I removed all
> >> >> >that crap that I've been adding for the past week and a half,
> >> >> >make that call and then call rmc_shutdown(), and not worry
> >> >> >about forks or any of that, because the vsvr id won't change,
> >> >> >at least not w/o a reboot.  This version has been running on
> >> >> >cougar-soak and seems to avoid the problem of pm not reaping
> >> >> >the sessions.  We no longer are seeing the errors from the
> >> >> >cluster daemon.
> >> >> >
> >> >> >
> >> >> >The files that are different from last time are dns.c,
> >> >> >smtp_out.c and exim.c. And of course
> >> >> >onstor-vsvr-support.dpatch itself.
> >> >> >
> >> >> >~andys/src/dev/linux/Pkgs/source/exim4/exim4-4.63-virgin/src
> >> >> >
> >> >> >~andys/src/r401rel/linux/Pkgs/source/exim4/exim4-4.63/build-tree/src
> >> >> >
> >> >> >On Tue, 11 Nov 2008 15:34:30 -0800 Maxim Kozlovsky
> >> >> ><maxim.kozlovsky@onstor.com> wrote:
> >> >> >
> >> >> >> Ok. Hope this is the last one.
> >> >> >>
> >> >> >> >-----Original Message-----
> >> >> >> >From: Andy Sharp
> >> >> >> >Sent: Tuesday, November 11, 2008 3:30 PM
> >> >> >> >To: Maxim Kozlovsky
> >> >> >> >Subject: Re: please review 31135
> >> >> >> >
> >> >> >> >On Mon, 10 Nov 2008 18:20:21 -0800 Maxim Kozlovsky
> >> >> >> ><maxim.kozlovsky@onstor.com> wrote:
> >> >> >> >
> >> >> >> >> Ok...
> >> >> >> >>
> >> >> >> >> The alarm and setitimer functions can not fail with EINTR
> >> >> >> >> according to my version of the man pages, but if it feels
> >> >> >> >> safer to call them twice or thrice, go ahead.
> >> >> >> >
> >> >> >> >OK, the only thing that's changed from yesterday is that
> >> >> >> >exim's sighandler function no longer installs itself as the
> >> >> >> >handler at the end any more, and no longer calls
> >> >> >> >rmc_timer_intr, because it doesn't need to anymore.
> >> >> >> >
> >> >> >> >> Fs-linux.c has some unrelated change at line 51.
> >> >> >> >
> >> >> >> >Aha, that snuck in because I did it on r401rel branch.
> >> >> >> >Fixed.
> >> >> >> >
> >> >> >> >
> >> >> >> >> >-----Original Message-----
> >> >> >> >> >From: Andy Sharp
> >> >> >> >> >Sent: Monday, November 10, 2008 5:45 PM
> >> >> >> >> >To: Maxim Kozlovsky
> >> >> >> >> >Subject: please review 31135
> >> >> >> >> >
> >> >> >> >> >The original and modified directories are:
> >> >> >> >> >
> >> >> >> >> >~andys/src/dev/linux/Pkgs/source/exim4/exim4-4.63-virgin/src
> >> >> >> >> >
> >> >> >> >> >~andys/src/r401rel/linux/Pkgs/source/exim4/exim4-4.63/build-
> >> >tree/src
> >> >> >> >> >
> >> >> >> >> >
> >> >> >> >> >
> >> >> >> >> >Change 31135 by andys@ripper on 2008/11/10 16:30:47
> >> >> >> >> >*pending*
> >> >> >> >> >
> >> >> >> >> >        TED00025710 - Over 200 Exim processes running
> >> >> >> >> >
> >> >> >> >> >        Try # 87 to fix this.
> >> >> >> >> >
> >> >> >> >> >        Nothing seemed able to deal with the SIGALRM
> >> >> >> >> > problem except daisy-chaining the RMC SIGALRM handler
> >> >> >> >> > to the tail end of the exim4 handler.  Bletch.
> >> >> >> >> >
> >> >> >> >> >        Also, fix up the exim config files to better
> >> >> >> >> > dev/null all local messages.  It actually says
> >> >> >> >> > => /dev/null in the log.  Cool.
> >> >> >> >> >
> >> >> >> >> >        Add fixups to exim-rm-frozen daily cron script
> >> >> >> >> > that better deals
> >> >> >> >> >	with the possible situation of an overload of
> >> >> >> >> >	frozen (undeliverable)
> >> >> >> >> >        or queued messages.
> >> >> >> >> >
> >> >> >> >> >        Change the /etc/hosts generating code to put the
> >> >> >> >> > hostname.ifname last, otherwise exim wants to treat the
> >> >> >> >> > local host as a remote host and do a DNS lookup on it,
> >> >> >> >> > which isn't helping.
> >> >> >> >> >
> >> >> >> >> >        reviewed by
> >> >> >> >> >
> >> >> >> >> >Affected files ...
> >> >> >> >> >
> >> >> >> >> >... //depot/dev/linux/Pkgs/binary/Deb/exim4-base_4.63-
> >> >17_mipsel.deb#6
> >> >> >> >> >edit ... //depot/dev/linux/Pkgs/binary/Deb/exim4-daemon-
> >> >custom_4.63-
> >> >> >> >> >17_mipsel.deb#9 edit
> >> >> >> >> >... //depot/dev/linux/Pkgs/source/exim4/onstor-vsvr-
> >> >support.dpatch#5
> >> >> >> >> >edit ... //depot/dev/linux/rootfs/etc/aliases#1 edit
> >> >> >> >> >... //depot/dev/linux/rootfs/etc/cron.daily/exim4-rm-frozen#3
> >> >> >> >> >edit ...
> >//depot/dev/linux/rootfs/etc/exim4/exim4.conf.template#4
> >> >> >> >> >edit ... //depot/dev/nfx-tree/code/ssc-genlib/fs-linux.c#10
> >> >> >> >> >edit
> >> >> >> >>
