IRC log of #zope for Tuesday, 2011-08-30

*** TomBlockley has joined #zope00:01
*** sp0cksbeard has left #zope00:37
*** TomBlockley has quit IRC00:53
*** J1m has quit IRC00:55
*** thetet has quit IRC01:03
*** sunew has quit IRC01:38
*** TresEquis has quit IRC01:47
*** m8 has quit IRC01:48
*** benji has quit IRC01:59
*** srichter has quit IRC02:03
*** alexpilz has joined #zope02:09
*** alexpilz1 has quit IRC02:11
*** alexpilz1 has joined #zope02:21
*** alexpilz has quit IRC02:23
*** srichter has joined #zope02:31
*** tiwula has quit IRC02:53
*** J1m has joined #zope02:56
*** _mup_ has quit IRC03:00
*** _mup_ has joined #zope03:00
*** J1m has quit IRC03:16
*** daMaestro has quit IRC03:16
*** ccomb has quit IRC03:18
*** runyaga has quit IRC03:25
*** Arfrever has quit IRC03:46
*** rbanffy has quit IRC03:49
*** dayne has joined #zope04:03
*** evilbungle has quit IRC04:19
dixondok folks - pop quiz; in RelStorage, the Oracle commit-lock-id - does this have to be an int, or can it be an arbitrary string?05:51
dixondwhat are the limitations/constraints?05:51
koshnot a clue at all05:51
koshI have not used Relstorage in production05:52
koshwhen I did testing on it with postgres I found it no faster then ZEO but more unreliable under high load so I decided not to use it05:52
koshrelstorage is faster then ZEO in 2.12 and under though by a lot05:52
koshbut ZEO in 2.13 got a MAJOR speed boost05:52
dixondyeah, unfortunately my driver is HA :-/  otherwise I'd just stick with ZEO - it's worked great for us for years.05:53
koshI had 4 zopes talking to ZEO/Relstorage and I found that when I I had about 1000 simul connections across the zopes that ZEO completed every tranasction05:53
koshI ended up with Relstorage and postgres not completing some transaction and if I made the test long enough postgres would just stop accepting all connections05:54
dixondLOL ouch.05:54
koshit would show no errors of any kind, it just would not accept any connections05:54
koshso I tried to use pgpool and pgool2 and both of them had the same problem05:54
koshand when I consulted docs on both of those that was a known issue05:54
dixondknow issue? holy hell, that's a doozy of a 'known issue' with no fix...05:55
koshso you had to install a heartbeat program that would autorestart pgpool or pgpool2 when they died since the devs had no idea why they died05:55
dixond'oops, I'm too busy'05:55
koshbut at least then you did not have to restart postgres05:55
dixondheartbeat sucks.05:55
koshso to replace ZEO with something that would be as reliable would involve postgres + pgpool2 + heartbeat and still not run as fast05:55
koshat that point I decided reliability did not really matter that much beyoned what ZEO did already05:56
koshI make backups once per day and could do once per hour or less if that was important and restore that in a few minutes05:56
koshI like the rackspace cloud servers05:56
dixondyeah. The dumb thing is I have *never* had ZEO die. Never. ... But it's all about 'perception' of risk in this scenario :-/05:56
koshyeah I have not had ZEO or Zope crash on me in years05:57
* dixond likes ZEO a lot05:57
koshthey are the most stable software I have worked with by a long shot05:57
koshand the security track record is amazing05:57
koshmost other systems have more exploits in a week then zope has had in the last 10 years combined!05:57
koshand the zope devs have taken reliability extremely seriously05:58
koshand with 2.13 ZEO got an insane speedup05:58
koshright now I have 2.13 ZEO and I am moving all my binary stuff to blobs for an even bigger gain05:58
koshafter I get all of that done I am going to test zlibstorage which is supposed to give another nice gain05:58
koshit just seems to match the reliability of ZEO with relstorage requires a lot of moving parts that all have to work correctly and that don't do that in reality05:59
dixondzlibstorage - zips everything by default inside ZODB?05:59
koshon a record by record basis05:59
koshit does not compress records that don't gain from compression05:59
dixondvery. cool.05:59
dixondless IO FTW05:59
koshand it turns out that since cpu time is so much vastly faster then disk io time that it really makes a speed difference05:59
koshsince you lower IO and trade it for a tiny ammount of cpu time06:00
dixondexactly. very nice.06:00
kosheven on a pentium 60 that was a good tradeoff (os/2 supported compressed exes for a huge gain)06:00
koshon a modern chip that is an insanely great tradeoff06:00
koshso you move all blobs to the filesystem and then compress all your remaining objects06:00
koshand you still have the insane reliability of zope + zeo06:01
dixondheh - just answered my question the old fashioned way: Try It And See   (Error: invalid literal for long(): shared-edit-main)06:01
koshI can see why stuff like postgres or mysql talk about HA, they need it!06:02
dixondyeah, they really do06:02
dixondZEO just never dies. it's a rock.06:02
koshI talked with one old friend recently and a mysql db they had needed about 4 hours worth of repairs every single day from corruption06:02
dixondwoah. that's crazy. holy hell.06:03
koshshe knew it was a problem but money was not there to fix it and the vendor told them to screw off06:03
koshthey paid a huge ammount of money for the app and the vendor would not support them at all06:03
koshand they had run out of money to get the app rebuilt06:03
koshI probably have not spent 4 hours recoving data from a corrupt ZEO in the last 4 years much less 4 hours per day06:04
dixondno. the only time I have ever spent with Zope was when I buggered a couple transactions up and truncated the DB to fix. And that was a no-brainer.06:06
koshapparently zope also does something completely novel and amazing compared to php before making a release06:09
koshthey actually run the unit test code to make sure it passes before releasing06:09
koshyou know that recent php security exploit for the crypto functions06:09
koshtheir test suite caught that problem along with a few hundred other problems06:09
koshbut they NEVER ran the test suite before making a release!06:09
dixondyou have got to be kidding. ugh.06:10
koshno that one is not a joke at all06:10
*** dayne has quit IRC07:04
*** Sada has joined #zope07:06
*** Rakshith has joined #zope07:07
*** kiran has joined #zope07:14
*** srichter has quit IRC07:19
*** Dikeshwar has joined #zope07:20
*** patul has joined #zope07:39
*** Bhavya_Ashwath has joined #zope07:58
*** dilip has joined #zope08:17
*** zhangkaizhao has joined #zope08:17
*** Allmity has joined #zope08:21
*** Gautam has joined #zope08:52
*** Gautam is now known as Guest3843408:52
*** __mac__ has joined #zope08:59
*** betabug-mb has joined #zope09:03
*** tisto has joined #zope09:04
*** slackrunner has joined #zope09:06
*** digitalmortician has joined #zope09:07
*** zhangkaizhao_ has joined #zope09:21
*** zhangkaizhao has quit IRC09:23
*** Wu has joined #zope09:23
*** goschtl has joined #zope09:27
*** slackrunner has quit IRC09:28
*** slackrunner has joined #zope09:31
*** agroszer has joined #zope09:35
*** slackrunner has quit IRC09:41
*** hever has joined #zope09:47
*** slackrunner has joined #zope09:48
*** hever has quit IRC09:49
*** do3cc has joined #zope09:50
*** hever has joined #zope09:51
*** alexpilz1 has quit IRC09:52
*** planetzopebot has quit IRC10:08
*** planetzopebot has joined #zope10:08
*** thetet has joined #zope10:37
*** zhangkaizhao_ has quit IRC10:41
*** zhangkaizhao has joined #zope10:42
*** alexpilz has joined #zope10:47
*** goschtl has quit IRC10:49
*** ccomb has joined #zope11:02
*** alga has joined #zope11:09
*** eperez has joined #zope11:09
*** sunew has joined #zope11:37
*** avoinea has joined #zope11:38
*** mr_jolly has quit IRC11:51
*** zhangkaizhao has quit IRC11:56
*** djc has joined #zope11:59
djcany hints on how that could happen?12:00
djc(when trying to run the test suite on zope.interface)12:01
*** TomBlockley has joined #zope12:01
djcit seems to get the installed zope.interface instead of the one being tested, but I don't get why12:01
bigkevmcddjc: what version is on your python path12:04
bigkevmcddjc: where are you running the tests from? what's in that directory?12:04
djcbigkevmcd: I'm running it inside Gentoo's packaging system12:06
djcsys.path looks like this: ['/var/tmp/portage/net-zope/zope-interface-3.7.0/work/zope.interface-3.7.0/src', '/var/tmp/portage/net-zope/zope-interface-3.7.0/work/zope.interface-3.7.0', '/home/djc/src/python-spidermonkey', '/home/djc/src/couchdb-python', '/var/tmp/portage/net-zope/zope-interface-3.7.0/work/zope.interface-3.7.0/build-2.7/lib.linux-x86_64-2.7', '/usr/lib64/', '/usr/lib64/python2.7', '/usr/lib64/python2.7/plat-linux2', '/usr/lib64/python2.12:06
bigkevmcdI don't know anything about Gentoo's packaging system, I haven't looked at Gentoo for a very, very long time12:06
*** vortec has joined #zope12:06
djcthen in that loop slightly before the failure, it picks up zope from the correct path, but it uses zope.interface from site-packages12:07
djcOBJ <module 'zope' from '/var/tmp/portage/net-zope/zope-interface-3.7.0/work/zope.interface-3.7.0/src/zope/'> PART interface12:07
*** sunew has quit IRC12:07
djcOBJ <module 'zope.interface' from '/usr/lib64/python2.7/site-packages/zope/interface/__init__.pyc'> PART tests12:07
vortechi, packing a Data.fs takes some time - is it safe to let zope running and let users make changes to objects in the meantime?12:08
*** mr_jolly has joined #zope12:08
*** zhangkaizhao has joined #zope12:09
*** sunew has joined #zope12:10
betabug-mbvortec: shouldn't make a difference12:14
*** mr_jolly has quit IRC12:15
vortecbetabug-mb: can you explain or do you have a link which explains what exactly happens when you pack a database?12:21
betabug-mbhmmm, it goes through the Data.fs and rebuilds it into a new one, using only the latest revisions of objects12:22
betabug-mbdunno if there is a better description out there12:22
*** mitchell`off is now known as mitchell`12:36
*** TomBlockley has quit IRC12:39
*** TomBlockley has joined #zope12:41
*** betabug-mb has quit IRC12:43
*** zhangkaizhao has quit IRC12:43
*** Wu has quit IRC12:52
*** TomBlockley has quit IRC12:53
*** sunew has quit IRC12:54
*** evilbungle has joined #zope12:55
*** teix has joined #zope12:55
*** evilbungle has quit IRC12:57
*** zhangkaizhao has joined #zope12:57
*** goschtl has joined #zope12:58
*** zhangkaizhao has quit IRC13:00
*** zhangkaizhao has joined #zope13:01
*** Sada has quit IRC13:03
*** Sada has joined #zope13:04
*** mr_jolly has joined #zope13:04
*** TomBlockley has joined #zope13:13
*** alexpilz1 has joined #zope13:14
*** alexpilz has quit IRC13:14
*** Wu has joined #zope13:16
*** slackrunner has quit IRC13:18
*** goschtl_ has joined #zope13:19
*** goschtl has quit IRC13:19
*** goschtl_ is now known as goschtl13:19
*** tisto is now known as tisto|away13:22
*** slackrunner has joined #zope13:23
*** evilbungle has joined #zope13:28
*** tisto|away is now known as tisto13:28
*** m8 has joined #zope13:48
*** dayne has joined #zope14:08
*** zhangkaizhao has quit IRC14:16
*** yvl has quit IRC14:18
*** fredvd has joined #zope14:27
*** benji has joined #zope14:28
*** slackrunner has quit IRC14:45
*** menesis has joined #zope14:47
*** sunew has joined #zope14:48
*** thetet has quit IRC14:54
*** Wu has quit IRC15:01
koshpacking a database is pretty much like a garbage collection run in java15:04
*** brguedes has joined #zope15:09
brguedeshi everybody15:09
brguedessomeone can tell me, if it is possible to get the zope version in command line15:10
brguedesfor zope 215:10
*** fredvd has quit IRC15:11
koshnothing is sprining to my mind on that one other then looking at the event.log file and seeing what version it puts in when it starts up15:21
*** fredvd has joined #zope15:22
brguedeskosh: ok thanls15:25
*** Allmity has quit IRC15:27
*** m8 has quit IRC15:33
*** J1m has joined #zope15:51
*** avoinea has quit IRC15:55
*** avoinea has joined #zope15:55
*** Allmity has joined #zope15:56
*** Allmity has quit IRC15:58
*** kiran has quit IRC16:00
*** brguedes has quit IRC16:00
*** ccomb has quit IRC16:11
*** goschtl has quit IRC16:12
*** eperez has quit IRC16:13
*** sp0cksbeard has joined #zope16:19
*** yvl has joined #zope16:21
*** yvl has quit IRC16:26
*** ajsmith has joined #zope16:26
*** yvl has joined #zope16:28
*** patul has quit IRC16:30
*** dilip has quit IRC16:31
*** hever has quit IRC16:32
*** Arfrever has joined #zope16:36
*** thetet has joined #zope16:36
*** tisto has quit IRC16:43
*** ccomb has joined #zope16:43
*** hever has joined #zope16:44
*** Dikeshwar has quit IRC16:46
dixondok folks, deep technical ZODB question coming16:47
dixondlet's say I have a ZODB with a bunch of stuff in it mount inside Zope as '/my-stuff'16:47
dixondnow lets say I change the zope.conf to make the mount point '/other-stuff'16:47
dixondmy observation is that the DB contents are no longer visible16:48
dixond1) is this expected?16:48
dixond2) is the only solution to use aliasing inside zope.conf, or is there some other way?16:48
*** dayne has quit IRC16:48
*** Bhavya_Ashwath has quit IRC16:49
*** goschtl has joined #zope16:49
*** Guest38434 has quit IRC16:54
*** tisto has joined #zope16:59
*** hever has quit IRC17:00
*** dayne has joined #zope17:05
*** dayne has quit IRC17:08
dixondConceptually, surely there is a way to 'cutnpaste' the content under the original mount path to a new monut path without having to use the aliasing juju17:11
*** rbanffy has joined #zope17:11
*** d2m has joined #zope17:13
*** Rakshith has quit IRC17:14
*** Wu has joined #zope17:15
*** yvl has quit IRC17:19
*** d2m has quit IRC17:26
*** goschtl_ has joined #zope17:26
*** goschtl_ has quit IRC17:27
*** goschtl has quit IRC17:28
dixondhmmm, I could possibly also open the storage manually and just rename the top-level (mountpoint) object?17:30
*** cwarner_ has quit IRC17:30
dixondhrm. it seems that mountpoints don't support the rename operation.17:34
*** tiwula has joined #zope17:35
*** thetet has quit IRC17:36
dixondeven more hmmm. why is 'meta_type' attribute required before cut/paste will work?17:36
dixondoh ugh.17:37
dixondthis is totally just about permissions.17:37
dixondso... reading the code looks like I can just do this 'by hand' as it were...17:39
*** cwarner has joined #zope17:41
dixondand just like magic, it worked.17:42
dixondall my data just magically appeared in the instance using the new mointpoint.17:43
dixondbest. thing. ever.17:43
*** djc has left #zope17:47
*** digitalmortician has quit IRC17:54
*** Sada has quit IRC17:55
*** cwarner has quit IRC18:01
*** hever has joined #zope18:01
*** cwarner has joined #zope18:19
*** slackrunner has joined #zope18:33
*** avoinea has quit IRC18:39
*** __mac__ has quit IRC18:41
*** avoinea has joined #zope18:42
*** digitalmortician has joined #zope18:52
*** mr_jolly has quit IRC18:54
*** mr_jolly has joined #zope18:54
*** mr_jolly has quit IRC18:57
*** mr_jolly has joined #zope18:57
*** daMaestro has joined #zope18:58
*** TomBlockley_ has joined #zope18:59
*** TomBlockley__ has joined #zope19:01
*** TomBlockley_ has quit IRC19:01
*** TomBlockley has quit IRC19:02
*** TomBlockley__ is now known as TomBlockley19:02
*** mr_jolly has quit IRC19:02
*** agroszer has quit IRC19:02
*** mr_jolly has joined #zope19:02
*** m8 has joined #zope19:05
*** Wu has quit IRC19:16
*** mr_jolly has quit IRC19:24
*** alga has quit IRC19:25
*** mr_jolly has joined #zope19:28
*** tisto has quit IRC19:28
*** mr_jolly has left #zope19:28
*** slackrunner_ has joined #zope19:39
*** digitalmortician has quit IRC19:40
*** slackrunner has quit IRC19:42
*** slackrunner_ is now known as slackrunner19:42
*** fredvd has quit IRC19:52
*** digitalmortician has joined #zope19:54
*** agroszer has joined #zope20:06
*** ARiKA_ has joined #zope20:10
*** alexpilz1 has quit IRC20:13
*** TomBlockley has quit IRC20:23
*** mitchell` is now known as mitchell`off20:26
*** avoinea has quit IRC20:26
*** menesis has quit IRC20:30
*** ARiKA_ has left #zope20:34
*** alga has joined #zope20:46
*** do3cc has quit IRC20:53
*** __mac__ has joined #zope20:54
*** vortec has left #zope20:56
*** agroszer has quit IRC21:02
*** ccomb has quit IRC21:04
planetzopebotLast Chance for Discounts on Training at Plone Conference 2011 (Plone News)
*** huajie has joined #zope21:16
*** rbanffy has quit IRC21:25
*** slackrunner has quit IRC21:28
koshhail weirdos!21:29
*** alecm has quit IRC21:47
*** teix has quit IRC21:55
*** alexpilz has joined #zope22:12
*** espent has joined #zope22:14
*** goschtl has joined #zope22:17
*** alexpilz1 has joined #zope22:20
*** alexpilz has quit IRC22:21
*** goschtl has quit IRC22:33
*** sunew has quit IRC22:36
*** fredvd has joined #zope22:40
*** fredvd|away has joined #zope22:40
*** fredvd has quit IRC22:45
*** __mac__ has quit IRC22:51
*** alecm has joined #zope23:04
*** hever has quit IRC23:09
*** rbanffy has joined #zope23:42

Generated by 2.15.1 by Marius Gedminas - find it at!