Custom Query (2152 matches)

Reported by nils.toedtmann, 3 years ago.

We should create a public AMI with CKAN pre-installed and configured such that users can easily create their own EC2 machine with a running CKAN to play with.

There are three phases:

1. [nils] Deploy an empty EC2 instance to become the CKAN image master instance
2. Install a CKAN and give it a standard configuration.
3. [nils] Create a AMI from the CKAN image master instance and publish it.

I am happy to do first and last. Who is installing and configuring CKAN?

Unfortunately AMIs are specific to region, architecture and storage type. We cannot maintain too many images, so a number of choices have to be made:

• Which distribution/version? Ubuntu 10.04 LTS
• Which architecture/instance-type? I suggest 64-bit/t1.micro
• Which region? I suggest us-east-1 and maybe eu-west-1
• Which storage type? EBS (way easier to make an AMI from than instance-store)
• Install CKAN from deb packages via mercurial/virtualenv? I assume the latter because the AMI is targeted to developers?

​1 ​2

Reported by nils.toedtmann, 3 years ago.

There is some evidence pointing to CKAN handling memory inefficiently or even leaking under certain conditions:

When we migrated ckan.net/thedatahub.org from eu7.okfn.org (32bit) to s053.okserver.org (64bit) (​ticket) we experienced extraordinary memory usage peaks (​ticket). Here are the observed value with Apache default settings:

• eu7, mpm-prefork: base level ~0.6GB, peaks up to 2GB
• s055, mpm-prefork: base level ~1GB, peaks up to 4GB
• s055, mpm-worker: base level ~1.5GB, peaks up to 6GB

William reduced the life-time of a WSGI CKAN process from 500 requests down to 25 requests (​changeset). This (together with two other tweaks) changed the situation drastically:

• s055, mpm-event: base level ~1.4GB, no peaks

This suggests that the more requests a CKAN processes serves over time, the more memory it consumes, aka bad memory management or a leak.

To prove this theory, one could reduce the total number of WSGI CKAN processes as much as possible without killing the performance (e.g. down to processes=3), and then observing the relation between maximum-requests=25...500 and memory consumption.

On 14/09/11 17:49, David Read wrote:

Someone to do a bit of top-down memory-use profiling would be very useful. Also useful would be something in the tests that reported what test cases use lots of memory - this could be in the nose plugin.

+1

Reported by nils.toedtmann, 3 years ago.

This is a scratch pad ticket with some comments on the current status of our ckan deb packages. I know that some of it is the deb packaging roadmap anyway, please forgive me if i mention them here again.

Rufus and me re-deployed some community ckan instances onto s022 (see ​http://trac.okfn.org/ticket/926). We followed the documentation ​http://docs.ckan.org/en/latest/install-from-package.html

• Deb package version number: the version of the deb package is "python-ckan 1309471251~149be76faabc+lucid-1", and it's hard to guess from there that it contains a ckan 1.4.2a
• When is 1.4.3/1.5.x expected as deb package?
• There was a bug in the DB upgrade script /usr/share/pyshared/ckan/migration/versions/029_version_groups.py (line 150) which looks like it was fixed 1.4.1==>1.4.2 but was nevertheless present in this deb package.
• The current script /usr/bin/ckan-std-install
  • does not set the Apache ServerName? according to the $INSTANCE variable
  • automatically configures a ckan extension named after $INSTANCE
  • depends on local postgres
  • could be replaced with "/usr/bin/ckan-deploy --name=ckan-std --domain=ckan-std.localhost (see next point)
• (i think this is exactly James' plan): have more generic deployment script /usr/bin/ckan-deploy as part of python-ckan which takes arguments like
  • --domain=cc.ckan.net
  • --aliases=$list-of-domains
  • --name=cc (defaults to "domain")
  • --no-db (does not configure a DB)
  • --sql-alchemy=$DB_CONFIG_STRING (also runs "paster --plugin ckan db upgrade --config")
  • --extension $list-of-extesions
  • ...