New Release Checklist

From genomewiki
Revision as of 20:48, 13 May 2010 by Rhead (talk | contribs) (hgcentral is pushed automatically now when you run buildHgCentralSql.csh)
Jump to navigationJump to search

Go to New Assembly Release Process Details page.

Start Complete Step Link to details doc
If the organism has an existing assembly, check to see if chroms have changed in size considerably since the last release of an assembly.


Check to see if anyone has given WUStL the heads-up so they can begin doing their N-Scan predictions. Send email to markd and Jeltje van Baren: jeltje at cse.wustl.edu.
Ask admins to add your assembly to the exclude list (both the gbdb and mysql rsync download targets) at: hgdownload:/opt/csw/etc/rsyncd.conf. This should be done before syncing /gbdb data to hgnfs1 (i.e., before testing on beta).
Complete staging of assembly on hgwbeta. details
Check assembly details on hgwbeta, including: BLAT/PCR, GenBank updates, .nib file, default position/tracks, orderKey, gateway page, joinerCheck, indices, makedoc, featureBits. details
Complete testing of all tracks in sub-pushQ
Warn Donna K. that the release is imminent so she can update docs. Read everything and test links in downloads, gateway, credits, currentGenomes. details
Send email to genome-mirror letting them know that there is a new assembly (and exactly how much data) and when it will be released for rsync. Give 24 hours notice. details
Adjust Release Log in main pushQ details


Check that all of the MySQL tables are in good repair: hgwbeta> sudo dbCheck.sh $db details
Request rsync of database: hgwbeta -> RR (email to push-request) details
Mark main pushQ entry as “push requested”
Add to genomeClade if needed (only if this is the first release for this organism (e.g. “aaaBbb1”)) details
Add line to genome-centdb.hgcentral.dbDb with active set = 0 (use checkMetaData.csh to get exact entry) details
Add the two lines to genome-centdb.hgcentral.blatServers details
Wait for pushes to RR to be complete
Turn on GenBank automatic updates on on the RR. Follow instructions for turning on updates on beta, but add assembly to rr.dbs instead of hgwbeta.dbs details
Test browser on RR by changing db= value in URLs details
Request push of downloads hgwdev -> hgdownloads: /usr/local/apache/htdocs/goldenPath/<db>/*. Verify that liftOver files are in /liftOver, NOT in /vsXXX directories. Remember to push pairwise alignments and liftovers (vs the new assembly) from other organisms' databases. details
Request dump & autodump of database (send email to cluster-admins to ask for dump to hgdownload into htdocs/goldenPath/$db/database/) details
Double check that all appropriate /gbdb files were pushed from dev -> hgnfs1 (during beta testing), including liftOver files. (i.e. compare /gbdb/$db/* dev vs. beta)
Create currentGenomes symlink and request push from hgwdev -> hgdownloads /usr/local/apache/htdocs/goldenPath/currentGenomes/<db_symlink> (this is for the FTP link) details
Confirm that hgdownloads, currentGenome, and FTP links are functional details
Change line in genome-centdb.hgcentral.dbDb to be active = 1 (this makes the assembly live on RR!) details
Enable liftOver on the RR Follow instructions for enabling liftOver on beta, but make changes to hgcentral (not hgcentralbeta): (on hgwdev)"hgsql -h genome-centdb" then in mysql type "use hgcentral" Test hgLiftOver and hgConvert immediately details
Complete testing in RR: tracks, PCR, Blat, liftOver, etc. (that required active=1).
Change or add line in genome-centdb.hgcentral.(and in hgcentraltest and hgcentralbeta).defaultDb to point to new DB and test details
Send email to Donna K. letting her know it has been pushed, and she will push all static documents. Test all of her files. Test links in downloads, gateway, credits, currentGenomes. details
For human, mouse and rat (KG assemblies), ask for push of htdocs/knownGeneList/$db/* and knownGeneLists.html
Ask admins to remove your assembly from the exclude list at: hgdownload:/etc/rsyncd.conf and alert genome-mirror.
If this is a new species, send an email to Branwyn (bwagman@ucsc.edu). There is no longer a place to announce new assemblies, but she said she wanted to be informed anyways in case she wants to write something about it.


Notify cluster-admin that new assembly is complete and ask for push to genome-mysql. Remind them to make permissions for users "genome" and "genomep", and update the mysql.db table permissions.
Push net and chain tracks in other organisms, taking care to edit and push trackDb as necessary. If these tracks are replacing nets and chains to a previous assembly, drop those tracks and notify the mirrors. details
Click on “done!” in main pushQ for this assembly (the release log will be updated the next morning)


Retire the assembly sub pushQ using retirePushQ.csh from hgwdev. Make sure there are release log entries for the net and chain tracks in other databases. (The script will remove release log notes for all push queue entries where dbs=$db.)
Update hgcentral.sql: cd $WEEKLYBLD; Run buildHgCentralSql.csh real (can run w/o 'real' to just see diffs). hgcentral.sql will be automatically pushed to hgdownload.
Double check next day: release log, database dump/autodump, genome-mysql, genbank auto-update, downloads (bigZip autodownloads) details