Mail Archive Search: <+subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*\[lvs\-users\]\s+DNS\s+LVS\s+questions\s*$/>

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*\[lvs\-users\]\s+DNS\s+LVS\s+questions\s*$/: 14 ]

Total 14 documents matching your query.

1. Re: [lvs-users] DNS LVS questions (score: 1): Author: Graeme Fowler <graeme@xxxxxxxxxxx>; Date: Thu, 27 Sep 2012 10:32:43 +0100; Thanks. I've forwarded that to one of the IPVS developers for their perusal. Graeme _______________________________________________ Please read the documentation before posting - it's available at: h; /html/lvs-users/2012-09/msg00031.html (10,399 bytes)
2. Re: [lvs-users] DNS LVS questions (score: 1): Author: Pete Ashdown <pashdown@xxxxxxxxxxxx>; Date: Wed, 26 Sep 2012 08:29:02 -0600; Ubuntu 12.04.1 linux-image-3.2.0-31-generic 3.2.0-31.50 ipvsadm 1:1.25.clean-1ubuntu5 ldirectord 1:3.9.2-5ubuntu4.1 bind-9.9.1-P3 with ratelimit patch sysctl.conf: net.ipv4.conf.default.rp_filter=1 n; /html/lvs-users/2012-09/msg00030.html (14,969 bytes)
3. Re: [lvs-users] DNS LVS questions (score: 1): Author: Graeme Fowler <graeme@xxxxxxxxxxx>; Date: Wed, 26 Sep 2012 09:21:46 +0100; I don't believe persistence is the issue here - it's the actual UDP timeout set using "ipvsadm --set <tcp> <tcpfin> <udp>" If Pete's system is hitting an OOM with defaults set, something is very wron; /html/lvs-users/2012-09/msg00029.html (10,933 bytes)
4. Re: [lvs-users] DNS LVS questions (score: 1): Author: Kelsey Cummings <kgc@xxxxxxxxxxxxxx>; Date: Wed, 26 Sep 2012 00:16:32 -0700; And, actually, I was looking in the wrong place. We have persistence turned off. Sorry! -K _______________________________________________ Please read the documentation before posting - it's availabl; /html/lvs-users/2012-09/msg00028.html (10,282 bytes)
5. Re: [lvs-users] DNS LVS questions (score: 1): Author: Pete Ashdown <pashdown@xxxxxxxxxxxx>; Date: Tue, 25 Sep 2012 12:48:16 -0600; ldirectord With the timeout at 1 second, I'm using less than 1.5GB RAM. It only OOMs on higher timeouts, like 15 seconds. _______________________________________________ Please read the documentation; /html/lvs-users/2012-09/msg00027.html (10,999 bytes)
6. Re: [lvs-users] DNS LVS questions (score: 1): Author: Kelsey Cummings <kgc@xxxxxxxxxxxxxx>; Date: Tue, 25 Sep 2012 14:11:57 -0700; We've never had any issues with that here, consequently we are using the default udp timeout of 30 seconds and have inactive connection counts in on each server floating around 17k. Centos 5.8, 2.6.1; /html/lvs-users/2012-09/msg00026.html (10,881 bytes)
7. Re: [lvs-users] DNS LVS questions (score: 1): Author: Kelsey Cummings <kgc@xxxxxxxxxxxxxx>; Date: Tue, 25 Sep 2012 11:29:18 -0700; ... Something isn't right there - I'm pushing >4k qps per lvs box in our dns clusters on hosts with 2GB, with 1.5G in cache. Keepalived? -- Kelsey Cummings - kgc@xxxxxxxxxxxxxx sonic.net, inc. System; /html/lvs-users/2012-09/msg00025.html (10,590 bytes)
8. Re: [lvs-users] DNS LVS questions (score: 1): Author: Pete Ashdown <pashdown@xxxxxxxxxxxx>; Date: Tue, 25 Sep 2012 08:42:10 -0600; 1000-2000 qps. 8GB of RAM. _______________________________________________ Please read the documentation before posting - it's available at: http://www.linuxvirtualserver.org/ LinuxVirtualServer.org; /html/lvs-users/2012-09/msg00024.html (11,438 bytes)
9. Re: [lvs-users] DNS LVS questions (score: 1): Author: Graeme Fowler <graeme@xxxxxxxxxxx>; Date: Tue, 25 Sep 2012 09:42:40 +0100; Really? How many queries/sec is your system handling? Those numbers are small enough to be practically insignificant. You're right though, in DR mode there's no record of an active connection as the; /html/lvs-users/2012-09/msg00023.html (10,812 bytes)
10. Re: [lvs-users] DNS LVS questions (score: 1): Author: Pete Ashdown <pashdown@xxxxxxxxxxxx>; Date: Sun, 23 Sep 2012 02:51:17 -0600; Thanks for the localhost addressing tip. I've got my lvs dns in production now and it is cranking along. I had to lower the udp timeout to 1 second from my predicted 15 because the latter was crashin; /html/lvs-users/2012-09/msg00021.html (9,986 bytes)
11. Re: [lvs-users] DNS LVS questions (score: 1): Author: Graeme Fowler <graeme@xxxxxxxxxxx>; Date: Thu, 06 Sep 2012 12:43:56 +0100; The easiest and most trivial way to resolve this is a requirement of the LVS-DR method - to have the VIP(s) bound to loopback interfaces as /32 addresses (as interface aliases, or using "ip address a; /html/lvs-users/2012-09/msg00003.html (9,496 bytes)
12. Re: [lvs-users] DNS LVS questions (score: 1): Author: Jesper Dangaard Brouer <brouer@xxxxxxxxxx>; Date: Thu, 06 Sep 2012 13:11:23 +0200; This is not caused by IPVS. The solution is that your DNS server must UDP bind/connect to the VIP, not the real-server IPs. The problem is caused by how UDP works in the kernel, and how the socket AP; /html/lvs-users/2012-09/msg00002.html (10,679 bytes)
13. Re: [lvs-users] DNS LVS questions (score: 1): Author: Kelsey Cummings <kgc@xxxxxxxxxxxxxx>; Date: Wed, 5 Sep 2012 12:39:09 -0700; There's a few ways to do this, but we have it setup like this: -A -t test-virt.ns.sr.sonic.net:mdns -s rr -a -t test-virt.ns.sr.sonic.net:mdns -r b.ns.sr.sonic.net:domain -g -w 1 -a -t test-virt.ns.s; /html/lvs-users/2012-09/msg00001.html (9,922 bytes)
14. [lvs-users] DNS LVS questions (score: 1): Author: Pete Ashdown <pashdown@xxxxxxxxxxxx>; Date: Tue, 04 Sep 2012 15:08:10 -0600; I'm trying to setup a DNS LVS like this: Server A: x.x.x.10 Server B: x.x.x.11 Server C: x.x.x.12 The VIPs are also in the same network x.x.x.2 and x.x.x.3, they are respectively assigned to A & B, o; /html/lvs-users/2012-09/msg00000.html (8,509 bytes)

Current List: 1 - 14
Page: [1]