Sunday, 2012-03-25

« Saturday, 2012-03-24 Index Monday, 2012-03-26 »
*** Ahmad_HP has quit IRC00:03
*** kindaopsdevy has quit IRC00:09
justinsb_Ahmad_HP: 2 things you could try... 1) Using config drive instead of cloud-init00:16
justinsb_and 2) if you collect the SQL queries & their timings I would be happy to see if there's any obvious DB problem00:16
*** danwent has quit IRC00:21
*** Ahmad_HP has joined #openstack-dev00:21
Ahmad_HP@justinsb_: Sorry for missing your chat00:23
Ahmad_HPjust read the email00:23
justinsb_Hi!00:23
justinsb_I think we should be able to have any database query return in <1 second00:23
Ahmad_HPYeah sure, I will try to collect them next week.00:23
justinsb_Even that would be pretty terrible, but at least it's a starting point00:24
justinsb_So if you're seeing something a lot worse, I'd like to fix it!00:24
Ahmad_HPBut even then, I think I am more tempted to use metadata information that is already present at instance creation time00:24
justinsb_The big advantage of the config drive is that it does exactly that!00:24
justinsb_And avoids all networking problems00:24
Ahmad_HPI haven't looked at instance creation methods in detail but I think more or less all the metadata information already exists there00:24
justinsb_What metadata information are you pulling down?  Just IP and SSH key?00:25
Ahmad_HPeverything00:26
Ahmad_HPnova-api.log.11:2012-03-11 23:44:59,950 INFO nova.api [-] 0.149s 10.6.10.168 GET /2009-04-04/meta-data/instance-id None:None 200 [Python-urllib/2.7] text/plain00:26
Ahmad_HP text/html00:26
Ahmad_HPnova-api.log.11:2012-03-11 23:45:03,002 INFO nova.api [-] 0.138s 10.6.23.250 GET /latest/meta-data/ami-launch-index None:None 200 [PycURL/7.19.5] text/plain te00:26
Ahmad_HPxt/html00:26
Ahmad_HPnova-api.log.11:2012-03-11 23:45:03,059 INFO nova.api [-] 0.96s 10.6.21.236 GET /latest/meta-data/ami-launch-index None:None 200 [PycURL/7.19.5] text/plain tex00:26
Ahmad_HPt/html00:26
Ahmad_HPnova-api.log.11:2012-03-11 23:45:03,060 INFO nova.api [-] 0.97s 10.6.8.178 GET /latest/meta-data/ami-launch-index None:None 200 [PycURL/7.19.5] text/plain text00:26
Ahmad_HP/html00:26
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,898 INFO nova.api [-] 10.569170s 10.6.21.44 GET /2009-04-04/meta-data/instance-id None:None 200 [Python-urllib/2.6] text/pl00:26
Ahmad_HPain text/html00:26
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,903 INFO nova.api [-] 0.113s 10.6.21.44 GET /2009-04-04/meta-data/ None:None 200 [Python-urllib/2.6] text/plain text/html00:26
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,905 INFO nova.api [-] 0.98s 10.6.21.44 GET /2009-04-04/meta-data/local-ipv4 None:None 200 [Python-urllib/2.6] text/plain te00:26
Ahmad_HPxt/html00:26
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,907 INFO nova.api [-] 0.94s 10.6.21.44 GET /2009-04-04/meta-data/reservation-id None:None 200 [Python-urllib/2.6] text/plai00:26
Ahmad_HPn text/html00:26
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,909 INFO nova.api [-] 0.96s 10.6.21.44 GET /2009-04-04/meta-data/local-hostname None:None 200 [Python-urllib/2.6] text/plai00:26
Ahmad_HPn text/html00:26
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,911 INFO nova.api [-] 0.96s 10.6.21.44 GET /2009-04-04/meta-data/placement/ None:None 200 [Python-urllib/2.6] text/plain te00:26
Ahmad_HPxt/html00:26
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,913 INFO nova.api [-] 0.96s 10.6.21.44 GET /2009-04-04/meta-data/placement/availability-zone None:None 200 [Python-urllib/200:27
Ahmad_HP.6] text/plain text/html00:27
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,915 INFO nova.api [-] 0.103s 10.6.21.44 GET /2009-04-04/meta-data/ami-launch-index None:None 200 [Python-urllib/2.6] text/p00:27
Ahmad_HPlain text/html00:27
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,917 INFO nova.api [-] 0.94s 10.6.21.44 GET /2009-04-04/meta-data/public-hostname None:None 200 [Python-urllib/2.6] text/pla00:27
Ahmad_HPin text/html00:27
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,919 INFO nova.api [-] 0.98s 10.6.21.44 GET /2009-04-04/meta-data/mpi/ None:None 200 [Python-urllib/2.6] text/plain text/htm00:27
Ahmad_HPl00:27
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,921 INFO nova.api [-] 0.96s 10.6.21.44 GET /2009-04-04/meta-data/mpi/az2 None:None 200 [Python-urllib/2.6] text/plain text/00:27
Ahmad_HPhtml00:27
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,923 INFO nova.api [-] 0.95s 10.6.21.44 GET /2009-04-04/meta-data/hostname None:None 200 [Python-urllib/2.6] text/plain text00:27
Ahmad_HP/html00:27
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,925 INFO nova.api [-] 0.99s 10.6.21.44 GET /2009-04-04/meta-data/ami-id None:None 200 [Python-urllib/2.6] text/plain text/h00:27
Ahmad_HPtml00:27
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,927 INFO nova.api [-] 0.97s 10.6.21.44 GET /2009-04-04/meta-data/public-keys/ None:None 200 [Python-urllib/2.6] text/plain00:27
Ahmad_HPtext/html00:27
justinsb_Ah ... OK.  Perhaps I should ask what metadata you really need? :-)  I don't doubt that cloud-init is pulling in everything!00:27
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,929 INFO nova.api [-] 0.95s 10.6.21.44 GET /2009-04-04/meta-data/public-keys/0/openssh-key None:None 200 [Python-urllib/2.600:27
Ahmad_HP] text/plain text/html00:27
Ahmad_HPnova-api.log.12:2012-03-10 21:24:43,931 INFO nova.api [-] 0.95s 10.6.21.44 GET /2009-04-04/meta-data/instance-action None:None 200 [Python-urllib/2.6] text/pla00:27
Ahmad_HPsorry for so much data. but yeah pretty much everything that cloud init asks for00:27
justinsb_paste.openstack.org is great, BTW00:27
Ahmad_HPThanks.  you suggested using config drive instead of cloud init00:28
Ahmad_HPbut I don't think we want to impose this limitation on images00:28
*** Shrews has joined #openstack-dev00:29
justinsb_Cloud-init is OK though? :-)00:29
justinsb_I understand though00:29
justinsb_You don't really want to switch!00:30
justinsb_It looks like each call is taking about 2 ms?00:30
Ahmad_HPmetadata retrieval is not so bad if the system is not loaded enough.....00:30
justinsb_Except the first, which takes 4 seconds (ish)00:30
*** dolphm has quit IRC00:31
Ahmad_HP4 sec for metadata?00:31
Ahmad_HPwish we could achieve that00:31
justinsb_Oh hold on... I'm confused here.  The timestamps & the durations don't seem to coincide.00:31
Ahmad_HP4 sec is when there is very less system load00:32
Ahmad_HPplease, don't look at the timestamp of above queies.....they are hitting memcache00:33
justinsb_I think I need to check out the log format in more detail00:33
justinsb_I think if you're able to collect the SQL queries that would be very helpful, if you think it is the DB00:34
Ahmad_HPsure00:34
Ahmad_HPthanks by the way00:34
justinsb_Ahmad_HP: Do you know if you have that patch that went in very late in Essex with all the DHCP requests?00:35
Ahmad_HPdon't think so.00:35
justinsb_I'm just trying to find the link00:36
Ahmad_HPstill, my suggestion is same. why don't we use the information that is is already fetched from db at insteance creation time.00:37
justinsb_Ahmad_HP: It's a sound suggestion, but it might not be the source of the slowdown!00:38
Ahmad_HPagree00:38
justinsb_I have to run ... I will look into this a little more00:39
*** roge has joined #openstack-dev00:39
Ahmad_HPsure, thanks00:39
justinsb_The SQL trace will be very helpful in narrowing down the problem00:39
*** maplebed has quit IRC00:40
*** maplebed has joined #openstack-dev00:40
*** novas0x2a|laptop has joined #openstack-dev00:48
*** Ahmad_HP has quit IRC00:50
*** dolphm has joined #openstack-dev00:55
*** dolphm has quit IRC01:10
*** dolphm has joined #openstack-dev01:18
*** rods has quit IRC01:23
*** asalkeld has joined #openstack-dev01:26
*** roge has quit IRC01:32
*** mdomsch has joined #openstack-dev01:52
*** roge has joined #openstack-dev01:57
*** dolphm has quit IRC02:10
*** dolphm has joined #openstack-dev02:14
*** mdomsch has quit IRC02:16
*** johnpostlethwait has quit IRC02:19
*** ncode has joined #openstack-dev02:21
*** ncode has joined #openstack-dev02:21
*** dolphm has quit IRC02:22
*** roge has quit IRC02:37
*** novas0x2a|laptop has quit IRC02:45
*** dolphm has joined #openstack-dev02:46
*** dtroyer is now known as dtroyer_zzz02:48
*** dtroyer_zzz is now known as dtroyer02:52
*** novas0x2a|laptop has joined #openstack-dev02:54
*** dolphm has quit IRC03:01
*** roge has joined #openstack-dev03:06
*** Shrews has quit IRC03:09
*** sleepsonzzz is now known as sleepsonthefloor03:12
*** dtroyer is now known as dtroyer_zzz03:33
*** ahale has quit IRC04:16
*** paulstallworth has quit IRC04:18
*** paulstallworth has joined #openstack-dev04:30
*** ywu_ has quit IRC04:39
*** novas0x2a|laptop has quit IRC04:44
*** novas0x2a|laptop has joined #openstack-dev04:53
*** Tesk0 has left #openstack-dev04:53
*** Tesk0 has quit IRC04:54
*** hattwick has quit IRC05:03
*** ahale has joined #openstack-dev05:21
*** littleidea has joined #openstack-dev05:44
*** reed has quit IRC05:58
*** ncode has quit IRC06:09
*** shevek_ has quit IRC06:15
*** novas0x2a|laptop has quit IRC06:40
*** paulstallworth has quit IRC06:43
*** novas0x2a|laptop has joined #openstack-dev06:59
*** rkukura has joined #openstack-dev07:30
*** bepernoot has joined #openstack-dev07:40
*** hattwick has joined #openstack-dev07:59
*** bepernoot has quit IRC08:27
*** zykes- has quit IRC08:43
*** novas0x2a|laptop has quit IRC08:44
*** relateable has joined #openstack-dev08:58
*** maploin has joined #openstack-dev09:21
*** maploin has joined #openstack-dev09:21
*** novas0x2a|laptop has joined #openstack-dev09:50
*** dachary has quit IRC10:29
*** maploin has quit IRC10:30
*** kyriakos has quit IRC10:37
*** dachary has joined #openstack-dev10:45
*** dachary has quit IRC10:49
*** zykes- has joined #openstack-dev11:11
*** dachary has joined #openstack-dev11:14
*** BruceZu has joined #openstack-dev11:20
*** asalkeld has quit IRC11:37
*** littleidea has quit IRC11:41
*** sleepsonthefloor is now known as sleepsonzzz12:08
*** LinuxJedi has quit IRC12:11
*** LinuxJedi has joined #openstack-dev12:15
*** LinuxJedi has quit IRC12:19
*** LinuxJedi has joined #openstack-dev12:20
*** dtroyer_zzz is now known as dtroyer13:43
*** dolphm has joined #openstack-dev13:45
*** koolhead17|away has joined #openstack-dev13:46
*** dolphm has quit IRC13:54
*** dtroyer is now known as dtroyer_zzz13:59
*** armaan has joined #openstack-dev14:00
*** ywu has joined #openstack-dev14:27
*** jcapel has quit IRC14:40
*** armaan has left #openstack-dev14:40
*** rods has joined #openstack-dev14:46
*** novas0x2a|laptop has quit IRC15:15
*** dachary has quit IRC15:21
*** dtroyer_zzz is now known as dtroyer15:28
*** dachary has joined #openstack-dev15:52
*** dachary has quit IRC16:21
*** armaan has joined #openstack-dev16:22
*** armaan has left #openstack-dev16:29
*** armaan has joined #openstack-dev16:51
*** armaan has left #openstack-dev16:56
*** LinuxJedi has quit IRC17:01
*** agonella has joined #openstack-dev17:33
*** ncode has joined #openstack-dev17:45
*** rkukura has quit IRC17:50
*** dachary has joined #openstack-dev17:51
*** ncode_ has joined #openstack-dev18:06
*** ncode has quit IRC18:09
*** ncode_ is now known as ncode18:09
*** anotherjesse_zz is now known as anotherjesse18:30
*** ncode has quit IRC18:50
*** anotherjesse is now known as anotherjesse_zz19:27
*** koolhead17|away is now known as koolhead1719:31
*** dtroyer is now known as dtroyer_zzz19:51
*** andrewsmedina has joined #openstack-dev20:07
*** zykes- has quit IRC20:08
*** zykes- has joined #openstack-dev20:09
*** ncode has joined #openstack-dev20:13
*** asalkeld has joined #openstack-dev20:26
*** maplebed has quit IRC20:28
*** maplebed has joined #openstack-dev20:29
*** ncode has quit IRC20:33
*** Transformer has joined #openstack-dev20:37
*** Transformer has quit IRC20:39
*** koolhead17 has quit IRC20:49
*** koolhead17 has joined #openstack-dev20:49
*** dwalleck has joined #openstack-dev20:59
*** dwalleck has quit IRC21:00
*** maplebed_ has joined #openstack-dev21:03
*** maplebed has quit IRC21:04
*** maplebed__ has joined #openstack-dev21:04
*** maplebed__ is now known as Guest563821:04
*** maplebed has joined #openstack-dev21:04
*** maplebed has quit IRC21:06
*** koolhead17 has quit IRC21:07
*** koolhead17 has joined #openstack-dev21:08
*** Guest5638 has quit IRC21:09
*** dolphm has joined #openstack-dev21:20
*** philr has joined #openstack-dev21:22
*** philr has quit IRC21:38
*** reed has joined #openstack-dev21:40
*** reed has quit IRC21:52
*** dolphm_ has joined #openstack-dev21:54
*** dolphm has quit IRC21:57
*** koolhead17 has quit IRC22:03
*** jog0 has joined #openstack-dev22:26
*** jog0 has left #openstack-dev22:26
*** jog0 has quit IRC22:27
*** pmyers has quit IRC22:39
*** pmyers has joined #openstack-dev22:39
*** thingee has quit IRC22:53
*** paulstallworth has joined #openstack-dev23:04
mikalI'm playing with novaclient, and it seems that "nova show <servername>" isn't working as expected. Is server name meant to just be the name show in the "name" field of the nova show output?23:15
mikalAhhh, its not in the format show in the output.23:28
Lumiereusually, nova show id is the way to go23:32
*** dolphm_ has quit IRC23:35
*** Mandell has joined #openstack-dev23:46
mikalDigging further, nova is showing displayname, but expecting hostname for the search term23:48
mikalWhich is a bit confusing23:48
mikalYeah, server name matters to me because most of my users are using euca tools, which I don't think displays id23:49
*** dachary has quit IRC23:55
« Saturday, 2012-03-24 Index Monday, 2012-03-26 »

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