Folks,
We've been having occasional issues with failover dhcpd. I went looking for "peer holds all free leases", and happened to run across https://bugzilla.redhat.com/show_bug.cgi?format=multiple&id=610219, which is rated important, and is supposed to be fixed in 3.0.5-24. Looking at the repo at kernel.org, all I see is what we have, dhclient-3.0.5-23.el5.x86_64.rpm.
Any idea when this update will be released?
mark "we won't mention 3.1 or 4.x"
On Thu, 2010-10-14 at 14:58 -0400, m.roth@5-cent.us wrote:
Folks,
We've been having occasional issues with failover dhcpd. I went looking for "peer holds all free leases", and happened to run across https://bugzilla.redhat.com/show_bug.cgi?format=multiple&id=610219, which is rated important, and is supposed to be fixed in 3.0.5-24. Looking at the repo at kernel.org, all I see is what we have, dhclient-3.0.5-23.el5.x86_64.rpm.
Any idea when this update will be released?
mark "we won't mention 3.1 or 4.x"
--- Bother to even look on Upstreams Site? It's not freely available yet as I see it.
John
JohnS wrote:
On Thu, 2010-10-14 at 14:58 -0400, m.roth@5-cent.us wrote:
Folks,
We've been having occasional issues with failover dhcpd. I went looking for "peer holds all free leases", and happened to run across https://bugzilla.redhat.com/show_bug.cgi?format=multiple&id=610219, which is rated important, and is supposed to be fixed in 3.0.5-24. Looking at the repo at kernel.org, all I see is what we have, dhclient-3.0.5-23.el5.x86_64.rpm.
Any idea when this update will be released?
Bother to even look on Upstreams Site? It's not freely available yet as I see it.
No, I hadn't. I'm just a tad surprised - that was rated "important", and looked as though it would be released soon. And with 6 coming soon, I was thinking, though I haven't gone to look, that they'd have 3.1 or 4.x.
Thanks, though.
mark
*Try find-out SPEC file and rpm source, recompile it, this is the faster way I think, if you have big issue.
Fedora they released 4.1, check it.
http://mirrors.isu.net.sa/pub/fedora/linux/releases/13/Everything/x86_64/os/...
*-------------------------- Best Wishes, Waleed Harbi
Dream | Do | Be
On Thu, Oct 14, 2010 at 11:44 PM, m.roth@5-cent.us wrote:
JohnS wrote:
On Thu, 2010-10-14 at 14:58 -0400, m.roth@5-cent.us wrote:
Folks,
We've been having occasional issues with failover dhcpd. I went looking for "peer holds all free leases", and happened to run across https://bugzilla.redhat.com/show_bug.cgi?format=multiple&id=610219, which is rated important, and is supposed to be fixed in 3.0.5-24. Looking at the repo at kernel.org, all I see is what we have, dhclient-3.0.5-23.el5.x86_64.rpm.
Any idea when this update will be released?
Bother to even look on Upstreams Site? It's not freely available yet as I see it.
No, I hadn't. I'm just a tad surprised - that was rated "important", and looked as though it would be released soon. And with 6 coming soon, I was thinking, though I haven't gone to look, that they'd have 3.1 or 4.x.
Thanks, though.
mark
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Howdy,
Yeah, this issue’s been brought up in the past. One of the maintainer of FC 13 made the srpm available (I believe he prepped the spec file) for RHEL/CentOS 5 for dhcpd 4.1. We started looking at it and the problems are the dependencies and the dependencies of the dependencies. There are no srpms or rpms for those and the project becomes a real monster.
We’ve found the problem is generally caused by the time being out of sync between the servers or the dhcpd daemon on one of the boxes dies. To mitigate this problem we’ve setup custom Nagios alerts which utilize dhcping to alert us when a dhcpd process dies. We’ve also setup custom scripts that will restart the dead process on the box. So, we’re just waiting for CentOS 6 to be released.
____________________________
Matt Ausmus
Network Administrator
Chapman University
635 West Palm Street
Orange, CA 92868
(714)628-2738
mausmus@chapman.edu mailto:mausmus@chapman.edu
"Nothing is ever accomplished by a reasonable man."
-Bucy’s Law
From: Waleed Harbi [mailto:waleed.harbi@gmail.com] Sent: Thursday, October 14, 2010 1:49 PM To: CentOS mailing list Subject: Re: [CentOS] dhcpd rpm
Try find-out SPEC file and rpm source, recompile it, this is the faster way I think, if you have big issue.
Fedora they released 4.1, check it.
http://mirrors.isu.net.sa/pub/fedora/linux/releases/13/Everything/x86_64/os/...
-------------------------- Best Wishes, Waleed Harbi
Dream | Do | Be
On Thu, Oct 14, 2010 at 11:44 PM, m.roth@5-cent.us wrote:
JohnS wrote:
On Thu, 2010-10-14 at 14:58 -0400, m.roth@5-cent.us wrote:
Folks,
We've been having occasional issues with failover dhcpd. I went looking for "peer holds all free leases", and happened to run across https://bugzilla.redhat.com/show_bug.cgi?format=multiple&id=610219, which is rated important, and is supposed to be fixed in 3.0.5-24. Looking at the repo at kernel.org, all I see is what we have, dhclient-3.0.5-23.el5.x86_64.rpm.
Any idea when this update will be released?
Bother to even look on Upstreams Site? It's not freely available yet as I see it.
No, I hadn't. I'm just a tad surprised - that was rated "important", and looked as though it would be released soon. And with 6 coming soon, I was thinking, though I haven't gone to look, that they'd have 3.1 or 4.x.
Thanks, though.
mark
_______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Mon, 18 Oct 2010, Ausmus, Matt wrote:
Yeah, this issue’s been brought up in the past. One of the maintainer of FC 13 made the srpm available (I believe he prepped the spec file) for RHEL/CentOS 5 for dhcpd 4.1. We started looking at it and the problems are the dependencies and the dependencies of the dependencies. There are no srpms or rpms for those and the project becomes a real monster.
When this was raised before, I looked and there was a dhcp-4.2.0-11.fc15.src.rpm in RawHide; today I find dhcp-4.2.0-12.fc15.src.rpm
As I recall from trial builds, the first required fairly invasive changes in the initscripts, and I was not willing to sacrifice the time on a test box to satisfy idle curousity
Running a build at the latest code I also find it dies on this file, probably a either a missed new interface, or a mistake in the .c code
gcc -DHAVE_CONFIG_H -I. -I../includes -I.. -DLOCALSTATEDIR='"/var"' -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic -fno-strict-aliasing -fPIE -D_GNU_SOURCE -I../bind/include -MT resolv.o -MD -MP -MF .deps/resolv.Tpo -c -o resolv.o resolv.c resolv.c: In function 'read_resolv_conf': resolv.c:52: error: 'O_CLOEXEC' undeclared (first use in this function) resolv.c:52: error: (Each undeclared identifier is reported only once resolv.c:52: error: for each function it appears in.) make[2]: *** [resolv.o] Error 1
It is not clear why local includes of the resolver libraries are present. Probably the right place to have these discussions are in Fedora or upstream, however
-- Russ herrold
Ausmus, Matt wrote, On 10/18/2010 01:11 PM:
Howdy,
<SNIP>
We’ve found the problem is generally caused by the time being out of sync between the servers or the dhcpd daemon on one of the boxes dies.
NTP does not keep them closely enough synchronized? OH, and in case you were not aware of it, you could run NTP on one of them using local clock if you don't have a good trust able time server available for some reason. Also making your DHCP machines NTP peers would be good too.
Or are you talking about some other type of time?
-----Original Message----- From: Todd Denniston [mailto:Todd.Denniston@tsb.cranrdte.navy.mil] Sent: Monday, October 18, 2010 10:43 AM To: CentOS mailing list Subject: Re: [CentOS] dhcpd rpm
Ausmus, Matt wrote, On 10/18/2010 01:11 PM:
Howdy,
<SNIP> > > We’ve found the problem is generally caused by the time being out of > sync between the servers or the dhcpd daemon on one of the boxes dies.
NTP does not keep them closely enough synchronized? OH, and in case you were not aware of it, you could run NTP on one of them using local clock if you don't have a good trust able time server available for some reason. Also making your DHCP machines NTP peers would be good too.
Or are you talking about some other type of time?
-- Todd Denniston Crane Division, Naval Surface Warfare Center (NSWC Crane) Harnessing the Power of Technology for the Warfighter
I think the problem we have/had was related to a Layer 8 issue. I have student workers build our DHCP servers from specifications that I give them. What more than likely happened was that dhcpdate wasn't run on the boxes when they were first built causing the date & time to be off and ntpd having issues getting the time back in line. My understanding of how ntpd works is that if the time is off too much it sees the time difference as irrational and is not able to pull it back in line. Once I run ntpdate on those boxes the issues go away and ntpd is able to maintain the time sync fine. ____________________________ Matt Ausmus Network Administrator Chapman University
"Nothing is ever accomplished by a reasonable man." -Bucy’s Law
-
Ausmus, Matt -
JohnS -
m.roth＠5-cent.us -
R P Herrold -
Todd Denniston -
Waleed Harbi