Page MenuHomePhabricator

Setup backup[12]01[456789] & backup[12]020 and migrate data to them; prepare for decommission backup[12]00[34567]
Closed, ResolvedPublic

Description

Setup:

  • backup1014
  • backup1015
  • backup1016
  • backup1017
  • backup1018
  • backup1019
  • backup1020
  • backup2014
  • backup2015
  • backup2016
  • backup2017
  • backup2018
  • backup2019
  • backup2020

Decommission:

  • backup1003
  • backup1004
  • backup1005
  • backup1006
  • backup1007
  • backup2003
  • backup2004
  • backup2005
  • backup2006
  • backup2007

Details

Related Changes in Gerrit:
SubjectAuthorRepoBranchLines +/-
Jcrespooperations/puppetproduction+0 -10
Jcrespooperations/puppetproduction+4 -36
Jcrespooperations/puppetproduction+41 -8
Jcrespooperations/puppetproduction+8 -4
Jcrespooperations/puppetproduction+1 -1
Jcrespooperations/puppetproduction+2 -4
Jcrespooperations/puppetproduction+62 -0
Jcrespooperations/puppetproduction+1 -0
Jcrespooperations/puppetproduction+5 -2
Jcrespooperations/puppetproduction+1 -1
Jcrespooperations/puppetproduction+1 -0
Jcrespooperations/puppetproduction+11 -2
Jcrespooperations/puppetproduction+3 -0
Jcrespooperations/puppetproduction+34 -11
Jcrespooperations/puppetproduction+2 -2
Jcrespooperations/puppetproduction+33 -14
Show related patches Customize query in gerrit

Event Timeline

Change #1254999 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackup: Deploy new 24 shards (6 hosts) for mediabackups@eqiad

https://gerrit.wikimedia.org/r/1254999

Change #1255003 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackups: Open s3 storage port on storage hosts from working hosts

https://gerrit.wikimedia.org/r/1255003

Change #1255000 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackup: Deploy new 24 shards (6 hosts) for mediabackups@codfw

https://gerrit.wikimedia.org/r/1255000

Change #1254999 merged by Jcrespo:

[operations/puppet@production] mediabackup: Deploy new 24 shards (6 hosts) for mediabackups@eqiad

https://gerrit.wikimedia.org/r/1254999

Change #1255828 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackup: Apply final role for eqiad mediabackup new storages

https://gerrit.wikimedia.org/r/1255828

Change #1255828 merged by Jcrespo:

[operations/puppet@production] mediabackup: Apply final role for eqiad mediabackup new storages

https://gerrit.wikimedia.org/r/1255828

Change #1255000 merged by Jcrespo:

[operations/puppet@production] mediabackup: Deploy new 24 shards (6 hosts) for mediabackups@codfw

https://gerrit.wikimedia.org/r/1255000

Change #1255877 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackups: Clarify iam directory

https://gerrit.wikimedia.org/r/1255877

Change #1255877 merged by Jcrespo:

[operations/puppet@production] mediabackups: Clarify iam directory

https://gerrit.wikimedia.org/r/1255877

Change #1255003 merged by Jcrespo:

[operations/puppet@production] mediabackups: Open s3 storage port on storage hosts from working hosts

https://gerrit.wikimedia.org/r/1255003

Change #1256082 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackup: Add missing parameter VGW_REGION corresponding to site

https://gerrit.wikimedia.org/r/1256082

Change #1256082 merged by Jcrespo:

[operations/puppet@production] mediabackup: Add missing parameter VGW_REGION corresponding to site

https://gerrit.wikimedia.org/r/1256082

Change #1256100 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackup: Followup bug fix to b3afe9d, fix syntax error

https://gerrit.wikimedia.org/r/1256100

Change #1256100 merged by Jcrespo:

[operations/puppet@production] mediabackup: Followup bug fix to b3afe9d, fix syntax error

https://gerrit.wikimedia.org/r/1256100

These old hosts cannot be decommissioned yet, they need to migrate its data to the new hosts first- they are still in production, just read only!

Change #1259901 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackup: Make the recovery account obsolete (but not remove it yet)

https://gerrit.wikimedia.org/r/1259901

Change #1259901 merged by Jcrespo:

[operations/puppet@production] mediabackup: Make the recovery account obsolete (but not remove it yet)

https://gerrit.wikimedia.org/r/1259901

Change #1261379 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackup: Add region parameter for rclone

https://gerrit.wikimedia.org/r/1261379

Change #1261379 merged by Jcrespo:

[operations/puppet@production] mediabackup: Add region parameter for rclone

https://gerrit.wikimedia.org/r/1261379

Change #1261387 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackup: Add resharding script using rclone

https://gerrit.wikimedia.org/r/1261387

I've started the data migration to the new hosts @ eqiad, emptying backup1004. Emptying a single host would take around ~16 days, maybe more. Other hosts can be done in parallel, to some extent, but ramping up the transfer speed very slowly.

Change #1261387 merged by Jcrespo:

[operations/puppet@production] mediabackup: Add resharding script using rclone

https://gerrit.wikimedia.org/r/1261387

Change #1266148 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] installserver: Treat any attempt to reimage backup hosts as an error

https://gerrit.wikimedia.org/r/1266148

Change #1266148 merged by Jcrespo:

[operations/puppet@production] installserver: Treat any attempt to reimage backup hosts as an error

https://gerrit.wikimedia.org/r/1266148

Change #1291904 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackup: Update s3cmd client configuration to eqiad/codfw sites

https://gerrit.wikimedia.org/r/1291904

Change #1291904 merged by Jcrespo:

[operations/puppet@production] mediabackup: Update s3cmd client configuration to eqiad/codfw sites

https://gerrit.wikimedia.org/r/1291904

Verification has finished, all files transferred ok. I found some metadata issues while doing the checking, but those do not block the decommissioning of the old hosts.

Change #1294929 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] mediabackup: Set backup[12]00[4-7] as insetup

https://gerrit.wikimedia.org/r/1294929

Change #1294929 merged by Jcrespo:

[operations/puppet@production] mediabackup: Set backup[12]00[4-7] as insetup

https://gerrit.wikimedia.org/r/1294929

Change #1294941 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] bacula: Setup backup1014 & backup2014 also as ES storage nodes

https://gerrit.wikimedia.org/r/1294941

Change #1294941 merged by Jcrespo:

[operations/puppet@production] bacula: Setup backup1014 & backup2014 also as ES storage nodes

https://gerrit.wikimedia.org/r/1294941

jcrespo added a subscriber: Marostegui.

Waiting until mid-August to decom 1003/2003, to ensure 90 day retention of ES db backups is never challenged CC @Marostegui .

Change #1320768 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] backup: Set backup1003 & backup2003 as insetup for decom.

https://gerrit.wikimedia.org/r/1320768

Change #1320768 merged by Jcrespo:

[operations/puppet@production] backup: Set backup1003 & backup2003 as insetup for decom.

https://gerrit.wikimedia.org/r/1320768

Change #1320928 had a related patch set uploaded (by Jcrespo; author: Jcrespo):

[operations/puppet@production] bacula: Remove last references to backup1003 & backup2003

https://gerrit.wikimedia.org/r/1320928

Change #1320928 merged by Jcrespo:

[operations/puppet@production] bacula: Remove last references to backup1003 & backup2003

https://gerrit.wikimedia.org/r/1320928

jcrespo updated the task description. (Show Details)

Last two hosts sent to decommissioning after reaching mid August buffer. Resolved.