Friday, 2013-10-11

*** vipul is now known as vipul-away00:04
*** openstackgerrit has quit IRC00:04
*** openstackgerrit has joined #openstack-dns00:04
*** ChanServ sets mode: +v openstackgerrit00:04
*** vipul-away is now known as vipul00:17
*** krow has quit IRC00:20
*** CaptTofu has quit IRC00:21
*** CaptTofu has joined #openstack-dns00:21
*** matsuhashi has joined #openstack-dns00:27
*** nosnos has joined #openstack-dns00:44
*** zane has joined #openstack-dns01:13
*** fandikurnia01 has joined #openstack-dns01:17
*** zane has quit IRC01:29
*** shakayumi has joined #openstack-dns02:06
*** ctracey|away has joined #openstack-dns02:17
*** krow has joined #openstack-dns02:28
*** msisk has joined #openstack-dns02:32
*** shakayumi has quit IRC02:39
*** krow has quit IRC02:41
*** zane has joined #openstack-dns02:42
*** zane has quit IRC02:44
*** artom has quit IRC02:44
*** zane has joined #openstack-dns02:44
*** artom has joined #openstack-dns02:44
*** shakayumi has joined #openstack-dns02:49
*** matsuhashi has quit IRC02:57
*** krow has joined #openstack-dns02:59
*** msisk has quit IRC03:13
*** msisk has joined #openstack-dns03:19
*** matsuhashi has joined #openstack-dns03:24
*** matsuhashi has quit IRC03:31
*** matsuhashi has joined #openstack-dns03:32
*** nosnos has quit IRC03:34
*** SlickNik has quit IRC03:34
*** SlickNik has joined #openstack-dns03:34
*** nosnos has joined #openstack-dns03:34
*** matsuhashi has quit IRC03:36
*** matsuhashi has joined #openstack-dns03:41
*** matsuhashi has quit IRC03:41
*** msisk has quit IRC03:42
*** matsuhashi has joined #openstack-dns03:42
*** matsuhashi has quit IRC03:46
*** CaptTofu has quit IRC03:47
*** CaptTofu has joined #openstack-dns03:48
*** matsuhashi has joined #openstack-dns03:49
*** krow has quit IRC04:19
*** krow has joined #openstack-dns04:35
*** krow has quit IRC04:40
*** krow has joined #openstack-dns04:51
*** fandikurnia01 has quit IRC05:00
*** fandikurnia01 has joined #openstack-dns05:00
*** fandikurnia01 has quit IRC05:20
*** fandikurnia01 has joined #openstack-dns05:24
*** zane has quit IRC05:25
*** krow has quit IRC05:41
*** krow has joined #openstack-dns05:50
*** krow has quit IRC05:52
*** krow has joined #openstack-dns06:14
*** fandikurnia01 has quit IRC06:57
*** fandikurnia01 has joined #openstack-dns06:58
*** dafter has joined #openstack-dns07:29
*** shakayumi has quit IRC07:39
*** ppenjoy has joined #openstack-dns09:22
*** ppenjoy has left #openstack-dns09:54
*** cflmarques has joined #openstack-dns10:17
*** CaptTofu has quit IRC11:27
*** CaptTofu has joined #openstack-dns11:27
*** fandikurnia01 has quit IRC11:42
*** krow has quit IRC11:49
*** krow has joined #openstack-dns11:52
*** betsy has joined #openstack-dns12:02
*** CaptTofu has quit IRC12:05
*** CaptTofu has joined #openstack-dns12:06
*** krow has quit IRC12:20
openstackgerritA change was merged to stackforge/designate: Update domains when servers are created, modified or deleted  https://review.openstack.org/4507812:21
*** fandikurnia01 has joined #openstack-dns12:28
*** matsuhashi has quit IRC12:33
*** matsuhashi has joined #openstack-dns12:34
*** nosnos has quit IRC12:35
*** nosnos has joined #openstack-dns12:35
*** matsuhashi has quit IRC12:38
*** dafter has quit IRC12:40
*** nosnos has quit IRC12:40
*** dafter has joined #openstack-dns12:41
*** dmakogon has joined #openstack-dns12:45
*** fandikurnia01 has quit IRC12:45
*** dafter has quit IRC12:46
*** fandikurnia01 has joined #openstack-dns12:46
*** krow has joined #openstack-dns12:57
*** CaptTofu has quit IRC13:00
*** CaptTofu has joined #openstack-dns13:00
*** fandikurnia01 has quit IRC13:04
*** vinodmr has joined #openstack-dns13:23
*** vinodmr has quit IRC13:33
*** dafter has joined #openstack-dns13:38
*** dafter has quit IRC13:38
*** dafter has joined #openstack-dns13:38
*** vinodmr has joined #openstack-dns13:40
*** eankutse has joined #openstack-dns13:45
cflmarquesHi Guys. Designate is fully functional with Bind9? Or, does it need some improvement? I m asking that because my companie have decided to use Bind instead of PowerDns.13:50
*** eankutse has quit IRC13:50
*** eankutse has joined #openstack-dns13:50
*** krow has quit IRC13:50
openstackgerritGraham Hayes proposed a change to stackforge/designate: Add status fields for domains and records  https://review.openstack.org/4473013:59
dmakogonCan anyone tell what is current status for incubation ?14:13
kiallcflmarques: heya, some of the RS folks are working on bettering the Bind9 support .. But, it's defiantly a second class citizen right now.14:15
kialldmakogon: we plan to revisit incubation after the summit14:15
*** msisk has joined #openstack-dns14:16
openstackgerritA change was merged to stackforge/designate: Add status fields for domains and records  https://review.openstack.org/4473014:22
cflmarquesKiall: "it's defiantly a second class citizen right now." ---> That's what I tried to explain here, but they insist on Bind. Why you guys decided to use PowerDns?14:25
artomcflmarques, in my case it's NSD, so I'm working on (and have submitted an initial patchset for review) on a standalone daemon that consumes Central's notifications and updates a slave NSD server.14:32
artomIt can be extended to support BIND.14:33
artomIf your company wants BIND, that could be your solution.14:33
artomUse the PowerDNS backend, but slave a BIND to it.14:33
artomWithout speaking for kiall, I suspect PowerDNS being MySQL powered made it the best choice for the backend.14:34
artomEasy updates.14:34
kiallcflmarques / artom: Yea - PowerDNS was certainly easier to integrate with, and fits our use case 100%.14:35
kiallAnd - PowerDNS is going to be much easier to build more advanced features with.. Like GeoIP/WRR/Failover etc14:36
*** jmcbride has joined #openstack-dns14:39
cflmarquesartom: Thank you for your advice! I will take a look on NSD, but I will try to convince people here to continue on PowerDns.14:39
cflmarques I agree that is much easier to build more advanced features whith it, but, in case I don't succeed I will try to follow your advice by slaving a BIND to the PowerDns Backed. It seem's a good solution.14:47
cflmarquesThank you Kiall / Artom14:48
*** zane has joined #openstack-dns14:51
*** vinodmr has quit IRC15:06
openstackgerritArtom Lifshitz proposed a change to stackforge/designate: Domain Import/Export  https://review.openstack.org/4955515:22
*** jmcbride1 has joined #openstack-dns15:24
*** jmcbride has quit IRC15:26
*** jmcbride has joined #openstack-dns15:31
*** jmcbride1 has quit IRC15:32
*** jmcbride1 has joined #openstack-dns15:32
*** jmcbride has quit IRC15:32
*** vinodmr has joined #openstack-dns15:33
*** jmcbride1 has quit IRC15:52
*** jmcbride has joined #openstack-dns15:56
*** vinodmr has quit IRC16:02
*** adrian_otto has joined #openstack-dns16:19
*** zane has quit IRC16:22
*** vinodmr has joined #openstack-dns16:28
mugsievinodmr: thanks for the review. put in answers for the questions in the wiki, and updating the patchset16:42
vinodmrthank you.16:42
mugsievinodmr: it is going to be a while, the backend notifications landed, so I need to move them around to deal with the pool manager....16:51
vinodmrmugsie, in the wiki, for (1) you mention "At that point, for most users I would imagine it would remain the same, unless they wanted to specify the pool for this domain".  So if the user, can specify the pool, would be changing the request to create a zone?16:54
kiallmugsie: we haven't talked about how zones get placed on pools, but I was thinking we follow nova's lead..17:01
kiallat create time, end users can provide a scheduler hint to decide which pool17:01
kiallotherwise, the scheduler follows it's default rules and places it on a "shared" / "public" pool17:01
kiallThen, a "migrate" call to move a zone from 1 pool to another where necessary17:02
vinodmrDoes the user initiate the migrate call or is it an admin?17:03
kiallIt would be a user thing I imagine, as the list of DNS servers would change as it moves to another pool17:04
vinodmrI agree17:07
vinodmrSo would there be a new "migrate" call then?17:07
mugsiekiall: yeah, that was what I was planing on, having the POST'd json have a pool_id if they want to override the default schedualer behavior17:12
mugsieI don' think it would be a migrate call.17:12
mugsiemaybe an update with a new pool-id?17:13
kiallvinodmr: eventually.. Migrate would be difficult, in that we'd need to keep the zone on the old pool for, say, 24 hours..17:13
kiallmugsie: maybe.. It's a while out :)17:13
mugsiehaving a migrate call breaks the REST principals17:13
*** jmcbride has quit IRC17:13
mugsieyeah, that is phase 3 ;)17:13
kiallmugsie: not really, If migrate is a resource where history of migrations is kept, it's totally valid REST17:15
vinodmrRegarding the authoritative name server - does every pool have an auth name server that is automatically inherited by any new zone in the pool?17:16
kiallThe canonical example of how to make actions on resources "RESTful" is the bank account money transfer example..17:16
kiallPOST /transfers with {"amount": 100.0, "from": "Graham", "to": "Kiall"}17:16
kiallvinodmr: yea, each pool would have its set of auth nameservers..17:18
kiallSo .. 1 pool might be the "shared" pool .. another might be Akamai nameservers, another might be designate managed nova instances running a DNS server...17:18
vinodmrSo would there be a default pool into which all zones without a pool hint would go to?17:20
kiallvinodmr: how familiar with nova are you? Giving Nova examples would be less typing for me :D17:21
mugsievinodmr: yes, or potentially multiple default pools, if you want to spread your load17:21
vinodmrNot much - but go ahead with the Nova examples and I will look them up17:21
kiallAt the end of the day, the "scheduler" code will have a set of rules, a set a pools, and will be asked to decide on the pool based on the zone being created and the rules it's been configured with17:22
kiallUsing Nova as an example..17:22
kiallThey have "Host Aggregates"17:22
kiallthese are groups of compute nodes17:22
kiallThe scheduler knows about them, and has some rules.. For example, windows instances go to 1 host aggregate, while linux goes to another, while flavors requiring GPU cards go to the GPU pool17:23
kiallDeployers are free to basically make any host aggregates and rules they need for their service17:24
kiallFor the DNS use case, lets say we have 2 set's of shared DNS servers.. Unicast non HA nameservers (cheap!) and Anycast nameservers spread all over the globe (expensive!)17:25
kiallThe scheduler might default to the cheap unicast pool, while end users could choose to more expensive Anycast option17:25
kiallOther users might create their own pool (we would boot up nova instances for them..)17:26
kiallOn your own pool, you get the entire namespace to yourself, along with possible access controls.. allowing you yo create "dev.local." and only allow your instances to query it17:26
kiallWhile another customer also has "dev.local."17:26
kiallIt's kinda a mix of Nova's Flavors and Host Aggregates, but the flavor model doesn't really seem to fit DNS in my mind :)17:27
vinodmrThat explanation is very helpful and clarifies things for me17:29
*** krow has joined #openstack-dns17:30
vinodmrWhen a user does a get zone/recordsets do we indicate the auth name server?17:30
vinodmrOr would there be a different way to get the auth name server in a pool?17:31
*** shakayumi has joined #openstack-dns17:52
*** alexb|f19 has joined #openstack-dns17:53
*** cflmarques has quit IRC17:57
*** krow has quit IRC17:59
*** jmcbride has joined #openstack-dns18:06
*** jmcbride has quit IRC18:07
*** jmcbride has joined #openstack-dns18:07
*** vinodmr has quit IRC18:09
kiallvinodmr: we have a /zones/ID/servers call which returns the list of nameservers18:14
kiall(It's domains/ID/servers in the current v1 API.. Obv..)18:14
kiallah .. he's gone18:15
*** CaptTofu has quit IRC18:27
*** CaptTofu has joined #openstack-dns18:28
*** vinodmr has joined #openstack-dns18:30
*** eankutse1 has joined #openstack-dns18:30
*** jmcbride has quit IRC18:32
*** CaptTofu has quit IRC18:32
*** eankutse has quit IRC18:32
vinodmrkiall:  With the /zones/ID/servers, it returns all the name servers - right.  Amongst these do we mark the name server in the SOA record differently?18:35
*** dafter has quit IRC18:35
*** jmcbride has joined #openstack-dns18:38
*** CaptTofu has joined #openstack-dns18:38
kiallToday /zones/ID/servers returns everything18:44
*** dafter has joined #openstack-dns18:44
kiallbut.. once pools are in, it will just return the nameservers for the pool belonging to that zone18:44
kiallCurrently, no.. We don't mark the nameserver listed in the SOA differently18:44
kiall(The end user shouldn't really care which server is listed in the SOA)18:45
artomShould that matter for zone import, actually?18:45
artom(Sorry, hijacking the convo)18:46
kiallNo, since the import will replace the SOA.. We only want the TTL and rname from the old SOA - I think18:46
*** alexb|f19 has quit IRC18:53
*** shakayumi has quit IRC18:53
vinodmrartom: regarding one of the comments in https://review.openstack.org/#/c/49555/3/designate/api/v2/controllers/zonefile.py, I wanted to find out if we do an export using these changes, can we pick up the contents that Designate exports and do an import on those, without any changes18:57
kiallvinodmr: yea, I've got that code cloned to my office desktop.. I was planning on trying that :)19:03
kiallA test case that does an export/import which includes every record type would be great :)19:04
*** ctracey|away is now known as ctracey19:06
*** eankutse has joined #openstack-dns19:07
*** eankutse1 has quit IRC19:07
artomYeah, import/export needs proper tests, beyond me importing all the zones we host, exporting them, and sorta kinda checking that nothing of value was lost.19:07
artom(Or gained!)19:08
artomBut to actually answer vinodmr's question, we can't.19:17
artomBecause the template used for export (same as writing zonefiles for the BIND backend) doesn't output $ORIGIN19:18
*** alexb|f19 has joined #openstack-dns19:21
vinodmrartom:  Wanted to check if this is correct - So import expects a $ORIGIN but export does not output it19:23
artomYeah.19:24
artomdnspython needs either an $ORIGIN in the zonefile, or passed as a parameter.19:25
artomThe bind9.jinja2 template doesn't print it.19:25
vinodmrJust wondering if in the code for import, can we figure out and pass a $ORIGIN parameter if the import contents does not have a $ORIGIN.19:28
*** jmcbride has quit IRC19:32
artomHrmm...19:42
artomOr change the template?19:42
*** adrian_otto has quit IRC20:13
*** dafter has quit IRC20:41
*** briancline has quit IRC20:44
*** tvb|afk has joined #openstack-dns20:44
*** vinodmr has quit IRC20:47
*** alexb|f19 has quit IRC21:04
*** CaptTofu has quit IRC21:17
*** CaptTofu has joined #openstack-dns21:17
*** eankutse has quit IRC21:18
*** CaptTofu has quit IRC21:22
*** eankutse has joined #openstack-dns21:22
*** vipul is now known as vipul-away21:28
*** vipul-away is now known as vipul21:28
*** eankutse has quit IRC21:33
*** jmcbride has joined #openstack-dns21:35
*** msisk has quit IRC21:37
*** adrian_otto has joined #openstack-dns21:38
*** adrian_otto has quit IRC21:40
*** adrian_otto has joined #openstack-dns21:41
*** vipul is now known as vipul-away21:43
openstackgerritArtom Lifshitz proposed a change to stackforge/designate: Contributing standalone tools  https://review.openstack.org/5069021:46
*** jmcbride has quit IRC21:50
*** vipul-away is now known as vipul21:54
*** zane has joined #openstack-dns22:09
*** tvb|afk has quit IRC22:14
*** CaptTofu has joined #openstack-dns22:30
*** CaptTofu has quit IRC22:39
*** CaptTofu has joined #openstack-dns22:40
*** adrian_otto1 has joined #openstack-dns22:42
*** adrian_otto has quit IRC22:44
*** CaptTofu has quit IRC22:44
*** adrian_otto1 has quit IRC22:46
*** jmcbride has joined #openstack-dns23:25
*** jmcbride has quit IRC23:38
*** jmcbride has joined #openstack-dns23:45
*** vipul is now known as vipul-away23:57
*** alexb|f19 has joined #openstack-dns23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!