Category: Development

Server update, 2016-07-04

This is a small bug-fix release. Server development has slowed down while we work on writing Docker containers for our new hosting infrastructure. But code contributions are always welcome. Thanks to Ulrich Klauer (chirlu) for his work on MBS-8806 this release.

The git tag is v-2016-07-04.

Bug

  • [MBS-8806] – Artist credits disappearing from tracklist & recordings after editing
  • [MBS-8987] – Untouched URLs are automatically/suddenly cleaned up
  • [MBS-8996] – Error message when trying to merge releases displays as “ARRAY(stuff)”
  • [MBS-8997] – Merging releases with “(Disc 1)”, “(Disc 2)” etc in titles selects all as disc 1
  • [MBS-8999] – Track lengths are in the wrong table column in the release editor

New Feature

  • [MBS-8994] – Allow passing &dismax=true through to the search server so Picard can show the same results as on the website

Server update, 2016-06-20

Our release today contains the usual lot of bug fixes and improvements. Thanks to yvanz, chirlu, and reosarevok for their contributions. The git tag is v-2016-06-20.

Bug

  • [MBS-7358] – Edit types not translated on Release editor / edit note tab
  • [MBS-7963] – Release editor doesn’t prevent too long disambiguation comments
  • [MBS-8809] – ASIN cleanup can be defective
  • [MBS-8951] – Recent notes is broken
  • [MBS-8968] – JSON version of artist includes gender-id value, which has no xml equivalent
  • [MBS-8977] – Edit searches with relationship type conditions cause an internal server error
  • [MBS-8983] – « Show only results that are in my subscribed entities. » checkbox broken
  • [MBS-8985] – Fix CPDL URLs matching and clean it up
  • [MBS-9000] – HTML special characters are not escaped on aliases page

Improvement

  • [MBS-7111] – Improve URL matching for musik-sammler.de links

Task

  • [MBS-8982] – Stop “fixing” piano with Guess Case

Server update, 2016-06-06

This server release contains fixes for bugs introduced in the recent schema change, and some other small improvements listed below. Thanks to reosarevok, chirlu, yvanz, and gcilou for their contributions this time around. The git tag is v-2016-06-06.

Bug

  • [MBS-7986] – Misleading error message for event setlists
  • [MBS-8941] – Beta: Last AC in a release sometimes doesn’t change with the rest when changing release artist
  • [MBS-8942] – Beta: cursor moves to the end of AC search field when removing a letter in the middle of the name
  • [MBS-8950] – iTunes favicon is not displayed on the sidebar
  • [MBS-8959] – Can’t add new packaging types – ISE

Improvement

  • [MBS-8824] – Update the rateyourmusic logo used in the sidebar
  • [MBS-8925] – Limit BookBrainz relationship to BookBrainz URLs

Task

  • [MBS-8957] – Update SecondHandSongs icon and add BBC logo in the sidebar

Updated search jar/war files

Given the utter slackers we are, we haven’t yet finished updating the search server to output the new MBIDs that were added to some entities in our last release. We’ll try and get that done soonish.

However, we did update the search code to fix this error in the search indexer:

ERROR: type “earth” does not exist

I’ve put both of these jar/war files on our FTP site:

If you would like to try and build these from source, you’ll need commit 4f677727 from mmd-schema and the latest master commit from search-server. For instructions on how to build this, please follow these instructions.

UPDATE: The build from the current master for search-server appears to not be able to load indexes upon startup. Please use the old war (we still use this in production) until we can release a fix.

Schema change release: What happened?

Now that we’ve finally finished the schema change release, I wanted to give an account of what happened in this arduous process. Before I dive into the details, I want to offer a picture that best sums up our current situation and challenges:

personal-container-mngmnt3

The shipping container is MusicBrainz and the boat is our hosting infrastructure. This picture perfectly describes the sort of challenges we’ve faced over the past few days. 🙂

Here is what happened:

Because the site was recently running slow and our search servers kept crashing, Zas and I were not available to help Bitmap prepare for the schema change release. This long process was left to Bitmap and Gentlecat to take care of on their own. We quickly realized that we were not ready for the release when the due date came and thus we delayed one week.

Sunday 22 May

Finally we were ready to proceed with the Postgres 9.5 upgrade. Once we started the process, we kept running into small problems that we didn’t get in our test setups. We do not have access to enough infrastructure to have a complete clone of our production environment, so we can only do so much to prepare for all the things that might happen when we run upgrades on our production servers.

All the while we attempted to start the upgrade, our backup database server was running much slower than anticipated. In the end we figured out that a step for optimizing the database (analyzing it) wasn’t carried out. During this time the site was really slow/unusable, but by the time the problem became apparent we had started the upgrade and could not turn back.

Once the upgrade was done, optimizing the database took much much longer than usual: 3 hours! This process wasn’t started until about 1am local time, which made for a very long night before that process finished. And even then we hit snags and had to start over a couple of times. At about 4:30am we had the site running on Postgres 9.5 in read only mode. The plan was to rest and start the schema change release in the morning.

Monday 23 May

Of course we had spent all of our time working on the Postgres upgrade and site stability, so our document that we use to plan the schema change was not in place. We spent the day preparing this and other bits for the release. To get an appreciation for what this document looks like, have a look! Note that some steps could be instant, others might take hours to carry out. Others might involve a sub-step or 20 not included in the document.

In the evening we were ready to make the change. By this point our backup DB was performing much better, so the read-only site worked acceptably. Thus, we started the release. Overall, the actual release process was reasonably smooth – we hit a few snags and had to do a lot of waiting for our slow servers. At about 1am in the morning things were finally complete. We proceeded with our sanity checks to make sure things went smoothly and all of them passed.

We proceeded to put the site into read-write mode and immediately saw portions of Postgres crashing, which is really bad. With community feedback we quickly deduced that some write operations were causing Postgres back-end processes to crash. We went back to read-only mode on the site and things stabilized and we finally went to bed at 3am.

Tuesday 24 May

In the morning we quickly found the source of database trouble with the help from the Postgres people on IRC. Thanks for the swift help Johto! We found that the steps for installing the updated third party extensions into Postgres had not completed correctly. Repeating the steps by hand fixed this problem.

Sadly yesterday morning we got an email informing us that our Live Data Feed replication stream had become corrupted. 🙁 This was heartbreaking news to us, since it means a great inconvenience to all of our Live Data Feed users. We immediately split into two teams: Zas, chirlu and myself to fix the root cause of the issue and Bitmap to investigate fixing the stream.

I proceeded to setup a test environment was able to quickly reproduce the problem. Zas and chirlu were an amazing support team Googling issues as I came across them. Within fairly short time we fixed the problem and deployed the fix to our database server. The problem was caused by a bug in a piece of code that we’ve been using for 13 years! A change in Postgres caused this bug to actually become a problem and corrupt our replication feed. 🙁

Once the problems were fixed we needed to initiate a new data dump and check to make sure the replication stream is working correctly. Of course we found a problem that we fixed and re-started the process to dump the data. Loads of hurry-up-and-wait situations to try our patience!

When we were satisfied that things were working correctly we re-enabled the site as read-write at about 1am and allowed people to continue editing. Exhausted we stumbled into bed waiting for data dumps to sync out to the FTP site.

Wednesday 25 May

Today Bitmap was flying home and as soon as WiFi became available on his flight he started working and helping with putting the schema change to bed. We’ve verified that everything is working as expected. At last this saga comes to and end and we can all take a break and catch up on sleep!

Thank you for your patience through all of this.

Schema change release, 2016-05-23 (with upgrade instructions)

Starting with this release, PostgreSQL 9.5 is now our minimum supported version. In order to import any future data sets, you will need to upgrade your installation to version 9.5.

Due to unforeseen problems with the Live Data Feed (AKA replication), users with slave databases will be required to first import a fresh data dump into their new 9.5 installation. We apologize that this is the case, but even had this stream not been broken, doing a clean import is faster and easier than doing the migration. For details on what happened during this rather lengthy schema change release, stay tuned for a post mortem blog post that covers the details.

If you have a non-replicated standalone database, you can use pg_upgrade and run ./upgrade.sh directly, but for simplicity we strongly recommend importing the latest data dump. Thus, we will only provide instructions for a clean import:

  1. Make sure you have PostgreSQL 9.5 installed, and your database settings in lib/DBDefs.pm are updated to point to the 9.5 installation if you currently have an older version of postgres running. If you already have postgres 9.5 and want to replace the existing database there, you’ll need to drop it first (using dropdb or from within psql). Be careful that you’re not dropping any important data if this is a standalone database that you’ve made changes to.
  2. Take down the web server running MusicBrainz, if you’re running a web server.
  3. Turn off cron jobs if you are automatically updating the database via cron jobs.
  4. Switch to the new code with git fetch origin followed by git checkout v-2016-05-23-schema-change-v2
  5. Run cpanm --installdeps --notest . to ensure your perl-based dependencies are up to date. Note the dot at the end.
  6. Set DB_SCHEMA_SEQUENCE to 23 in lib/DBDefs.pm
  7. Download the latest data dumps. If you don’t need historical edit data, excluding the edit dump will speed up your import significantly.
  8. Initialize a new database from the data dumps downloaded in step 7. Detailed instructions for doing this are located in INSTALL.md in the musicbrainz-server repository; if your data dumps are in /tmp, the command should simply be something like ./admin/InitDb.pl --createdb --import /tmp/mbdump*.tar.bz2.
  9. After the import has finished, turn cron jobs back on, if applicable.
  10. Restart the MusicBrainz web server, as well as memcached, if applicable.

We would like to thank bitmap, Gentlecat, zas, chirlu, reosarevok, gcilou for contributing directly to the release and we’d also like to thank all of the people who helped test, debug or otherwise offer support in this quite difficult release. Thank you!

And finally, here’s the list of changes you can expect in the upgrade:

Bug

  • [MBS-6406] – Admins can’t change email addresses
  • [MBS-8288] – Missing indexes for inverse lookup on *_gid_redirect tables
  • [MBS-8669] – Primary key for place table missing on old slaves
  • [MBS-8906] – Release pages ISE if CB doesn’t return JSON from its API for whatever reason
  • [MBS-8928] – If you submit the release editor without being logged in, it displays “[object Object]” as an error mesage
  • [MBS-8943] – Some pages do not respect DB_READ_ONLY setting

Improvement

  • [MBS-1873] – Fix vote tallies for edits
  • [MBS-3887] – Duplicate artist and label names not being checked against alias
  • [MBS-8287] – Log deleted entities that were in a subscribed collection
  • [MBS-8433] – Work attributes don’t have a uuid
  • [MBS-8716] – Store the edit data in a JSONB column
  • [MBS-8717] – Move the edit data to a separate table
  • [MBS-8838] – Add gids to all *_type* tables
  • [MBS-8873] – Convert and unify artist credit editors to React
  • [MBS-8909] – Add logos to IMDb and VGMdb links in the sidebar
  • [MBS-8939] – Update the Instagram logo used in the sidebar
  • [MBS-8940] – Let banner message editors dismiss the banner only temporarily

Task

  • [MBS-8656] – Bring edit table indexes back into sync
  • [MBS-8719] – Stop materializing of edit and vote counts
  • [MBS-8720] – Add a materialized view of edit note recipients
  • [MBS-8727] – Prevent duplicate votes
  • [MBS-8800] – Create the earthdistance extension and add a geodetic index for place coordinates
  • [MBS-8804] – Add BRIN indexes for timestamp columns
  • [MBS-8897] – add new entity icons
  • [MBS-8938] – Schema changes to support alternative tracklists

Schema change update

We’ve finally completed the schema update and things are returning to normal. We need to get a new data dump out and then we will provide upgrade instructions tomorrow. As you might be able to guess, unless you are already on Postgres 9.5, we are going to recommend a clean data import, rather than a migration, if you have a replicated slave.

And, if anyone even dare ask (within the next week) when an updated VM will be released, you owe the whole development team each 2 bars of high quality chocolate.

Feeling lucky, punk?

P.S. Can you tell we’ve been up too long? 🙂

Server capacity update

Zas and I have been working hard to improve the capacity and stability of the site. In the last week, we’ve identified and fixed at least 3 problems with the search servers and we’ve added a timeout function that times out queries that take longer than 3 seconds. We think that the main cause of trouble was that queries were piling up after a slow query ran too long and that the servers never recovered from that and consequently crashed.

We won’t go as far as saying that the search servers are fixed — every time we have a smidgen of hope that things are improving, they crash again. Seemingly out of spite! So, the search servers are better. 😉

Zas has also made a number of changes to the gateways and how we rate limit our incoming traffic. The rate limiting is now being done in a smarter way that reduces the overall traffic on our web servers. Well done!

We’ve also increased our bandwidth budget by 4mbits per second, which makes the site feel considerably more responsive.

Let me put these improvement into numbers: About a week ago were were struggling to keep up 250 requests per second and the site felt very sluggish. Now we can handle 500 requests a second and the site feels considerably faster. For large chunks of the day we are managing to handle all the traffic we should handle. And, the search servers haven’t crashed in 4 days!

We hope that this will give us a solid base from which to release the scheme upgrade tomorrow. Then once that is complete, we will start work on moving to the new hosting company.

Thanks for being patient with us!

Help! Is there a Lucene doctor in the house?

UPDATE: Thanks to user selckin in the #lucene IRC channel for quickly solving this for us! Hopefully we can put this fix into production later today!

As our regular readers may know, we’ve been having lots of troubles with our lucene based search servers. Over the past few days we’ve spent a fair amount of time, tuning, debugging and otherwise trying to troubleshoot our setup. We’ve fixed and identified a number of problems, but most importantly we feel that we’ve identified the core issue: Our servers are simply overloaded.

Under normal conditions we find our servers loaded to about 25% – 35% CPU — things look good and we don’t think we have a capacity problem with our servers. Then a slow query comes in that starts to slow things down. Much like a traffic jam that evolves out of thin air, one slow query can make a giant mess for everyone.

We’ve started timing our queries and most of the time, they can be measured in milliseconds. However, when things get bad, they may take up to 7-8 seconds. Our upstream web servers time out on the search request after about 5 seconds in order to prevent traffic from getting backed-up. What we need to do next is to limit the duration that a lucene query can run and terminate it after the timeout.

I’ve started looking at this and quickly realized that this is much more of a job than adding a simple timeout parameter to the search call. We’re currently using this search function from IndexSearcher:

  public TopDocs search(Query query,  int n);

Ideally I would like to add a way to timeout queries after 3 seconds. So far, I’ve discovered that we could use

  public void search(Query query, Collector results)

with a TimeLimitedCollector. The old call returns TopDocs and our code assumes that we have a TopDocs object from which to cull our search results. Having stared at the docs for lucene for a while, I haven’t found an way to convert the data in TimeLimitedCollector and convert it to TopDocs. It doesn’t make sense to me. 🙁

How does one do this? Sadly, we have no Java programmers on our team, so we’re quite a bit out of our league here. Is there an easier way to do this? Would someone be willing to write this code for us and submit a PR? We’d find some really good chocolate and send it to you if you do!

More info on our project:

We are using Lucene 4.10.4 on a custom codebase that pre-dates SOLR — we have a new SOLR project to replace this one, but it isn’t quite done yet. (Again, not having Java programmers is a bit of a problem for us).

Any tips, explanations or pull requests would be deeply appreciated! Chocolate reward offered!

Thank you!