| kraut | moin | 08:06 |
|---|---|---|
| ivoks | hi | 09:21 |
| thegodfather | hi ivoks | 09:21 |
| ivoks | is there any chance to include drbd module into ubuntu kernel? | 09:21 |
| ivoks | thegodfather: hi :) | 09:21 |
| thegodfather | ivoks: i did add a new patch to the bug.. any luck with that one? | 09:22 |
| ivoks | thegodfather: i have to wait night to test that patch :) | 09:22 |
| thegodfather | ivoks: ok cool. thanks | 09:22 |
| thegodfather | ivoks: drbd isn't exactly stable and portable.. last time i checked at least | 09:22 |
| ivoks | not stable? | 09:23 |
| ivoks | hm... :) | 09:23 |
| thegodfather | well it didn't use to be | 09:23 |
| thegodfather | and didn't build on many arches | 09:23 |
| thegodfather | ivoks: i think i might know why you were not able to reproduce the problem on vmware too | 09:24 |
| thegodfather | there is a specific algo to select the source address of an outgoing connection | 09:24 |
| ivoks | you do? i was thinking that it could be driver related stuff :/ | 09:24 |
| thegodfather | when the IP's are all in the same subnet | 09:25 |
| thegodfather | it matters only the last octet | 09:25 |
| thegodfather | let say: | 09:25 |
| thegodfather | node1: x.y.z.101 | 09:25 |
| thegodfather | node2: x.y.z.102 | 09:25 |
| thegodfather | node3: x.y.z.103 | 09:25 |
| thegodfather | etc. | 09:25 |
| thegodfather | and you slam the VIP's in the range from 1 to 100 | 09:25 |
| thegodfather | i don't remember if the lower is privileged | 09:25 |
| thegodfather | or the higher is | 09:25 |
| thegodfather | so perhaps you want to compare the real ip's on the running cluster | 09:26 |
| thegodfather | with the vmware setup | 09:26 |
| thegodfather | if you understand what I mean | 09:26 |
| thegodfather | perhaps on the running cluster the VIPs are below or up the real IP's | 09:26 |
| thegodfather | and in the vmware setup they are inverted | 09:26 |
| ivoks | on these nodes i have two addess on every node | 09:26 |
| ivoks | one is private and one is public | 09:26 |
| ivoks | so, one range is 192.168.0.x, and the other is xx.xx.xx.x | 09:27 |
| hytham__ | hi every one ...can some one help me in this issue (after compinilng a new kernel i recievd the following error "run-init:nuking initramfs contents:directory not empty" and kernel panics ??? | 09:27 |
| ivoks | i tried with the same setup on vmware, but all ips were from the private range | 09:27 |
| thegodfather | ivoks: right.. that can change a lot | 09:27 |
| thegodfather | because of the selection algo | 09:27 |
| ivoks | yeah, probably | 09:27 |
| thegodfather | do you have a chance to test the exact copy on vmware? | 09:28 |
| thegodfather | including real ips? | 09:28 |
| ivoks | not really at the moment | 09:29 |
| thegodfather | if so.. you might be able to reproduce the issue without a patched kernel | 09:29 |
| thegodfather | ok | 09:29 |
| ivoks | i'll test the patch on real systems tonight | 09:29 |
| thegodfather | ok | 09:29 |
| thegodfather | it was just to avoid you any more troubles | 09:29 |
| thegodfather | with the production system | 09:29 |
| thegodfather | ivoks: otherwise is there a way i can get the real config with real ip's? | 09:30 |
| ivoks | well, system is still in alpha stage, so i can get along with it :D | 09:30 |
| thegodfather | of course as NDA or something | 09:30 |
| thegodfather | that way i can set it up here in my vmware environment | 09:30 |
| thegodfather | ok | 09:30 |
| thegodfather | that case.. it's all good :) | 09:30 |
| ivoks | it the patch doesn't work, i think we can't talk about providing real config | 09:30 |
| ivoks | sorry... can | 09:31 |
| ivoks | not can't :) | 09:31 |
| thegodfather | ok | 09:31 |
| thegodfather | take into account that the new patch sucks in all changes in lowcomms.c that have been done to fix issues up till today | 09:31 |
| thegodfather | if this isn't working, we are probably experiencing another bug that we think | 09:31 |
| thegodfather | so that will require much more debugging | 09:31 |
| ivoks | well, i'll do everything i can to help you out | 09:32 |
| thegodfather | i really appreciate | 09:32 |
| thegodfather | thanks | 09:32 |
| ivoks | it's in my intrest to make it work too | 09:32 |
| thegodfather | yeps.. | 09:33 |
| === ivoks_ is now known as ivoks | ||
| ivoks | thegodfather: cross your fingers :) | 10:31 |
| * thegodfather crosses his fingers... | 10:31 | |
| ivoks | i hate rebooting system in testing kernel | 10:34 |
| ivoks | specially when they are thousands of kilometers away :) | 10:34 |
| ivoks | they're up! | 10:37 |
| ivoks | and printk shows stuff gets executed | 10:37 |
| ivoks | thegodfather: :* works for me :D | 10:41 |
| thegodfather | ivoks: rock on | 10:43 |
| thegodfather | ivoks: please report it in the bug report | 10:43 |
| ivoks | i'm doing last test just to be sure... | 10:43 |
| ivoks | will do, of course... | 10:43 |
| ivoks | yes, it works... kudos to thegodfather :) | 10:48 |
| thegodfather | perfect | 10:51 |
| thegodfather | thanks for testing man | 10:51 |
| ivoks | np | 10:52 |
| ivoks | thegodfather: funny thing... bringing old kernel back results in oopsing :) | 11:16 |
| thegodfather | well no idea.. sorry | 11:18 |
| ivoks | dlm_lock crashes on one node, not both... well, never mind, if patched one works, that's all i need :) | 11:18 |
| === chuck is now known as zul | ||
| === calc_ is now known as calc | ||
| === TheMuso_ is now known as TheMuso | ||
| === asac_ is now known as asac | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!