Page MenuHomePhabricator

Migrate some s4 hosts to file per table
Closed, ResolvedPublic

Description

These "new" s4 hosts (and probably the rest too) were probably recloned from a host not using file per table and they have a huge ibdata file and a few ibd tables:

root@db1081:/srv/sqldata/commonswiki# ls -lh | grep ibd
-rw-rw---- 1 mysql mysql 9.0M Mar 22 06:53 babel.ibd
-rw-rw---- 1 mysql mysql 9.2G Mar 22 09:53 filearchive.ibd
-rw-rw---- 1 mysql mysql 128K Mar  9 21:11 linter.ibd
-rw-rw---- 1 mysql mysql  18G Mar 22 09:55 page.ibd
-rw-rw---- 1 mysql mysql 106G Mar 22 09:55 revision.ibd
-rw-rw---- 1 mysql mysql 112K Dec  8 23:50 securepoll_elections.ibd
-rw-rw---- 1 mysql mysql 8.0M Mar 22 05:49 user_groups.ibd
root@db1081:/srv/sqldata# ls -lh ibdata1
-rw-rw---- 1 mysql mysql 1.2T Mar 22 09:56 ibdata1

Servers that need to be migrated to file per table:

  • db1059 (no longer in s4)
  • db1084
  • db1091
  • db1097

Details

ProjectBranchLines +/-Subject
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+3 -3
operations/mediawiki-configmaster+1 -1
operations/softwaremaster+1 -1
operations/mediawiki-configmaster+3 -3
operations/puppetproduction+4 -4
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+2 -2
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
operations/mediawiki-configmaster+1 -1
Show related patches Customize query in gerrit

Event Timeline

There are a very large number of changes, so older changes are hidden. Show Older Changes

This task is not strictly blocking: T17441 but it would be nice to get it done at some point

Change 346503 had a related patch set uploaded (by Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Depool db1081

https://gerrit.wikimedia.org/r/346503

Change 346503 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Depool db1081

https://gerrit.wikimedia.org/r/346503

Mentioned in SAL (#wikimedia-operations) [2017-04-05T06:55:03Z] <marostegui@tin> Synchronized wmf-config/db-codfw.php: Depool db1081 - T161088 (duration: 00m 39s)

Mentioned in SAL (#wikimedia-operations) [2017-04-05T06:55:54Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Depool db1081 - T161088 (duration: 00m 39s)

Mentioned in SAL (#wikimedia-operations) [2017-04-05T06:56:41Z] <marostegui> Stop replication on db1081 for maintenance - T161088

I have started to defragment db1081, in order to have at least have one healthy host with file per table to be able to reclone the new eqiad servers that arrived while they get racked.
Note: I will not reimage it, as that means to get 10.0.30 - better to leave the hosts we reimaged on s5 running for a few more days/weeks before we go for it.

Mentioned in SAL (#wikimedia-operations) [2017-04-06T06:02:37Z] <marostegui> Configure and start replication on db1081 after the defragment - T161088

db1081 has been defragmented and migrated to file per table. It is now catching up.

Change 346694 had a related patch set uploaded (by Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Repool db1081 with low weight

https://gerrit.wikimedia.org/r/346694

Change 346694 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Repool db1081 with low weight

https://gerrit.wikimedia.org/r/346694

Mentioned in SAL (#wikimedia-operations) [2017-04-06T07:16:22Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Repool db1081 with low weight - T161088 (duration: 00m 48s)

Change 346701 had a related patch set uploaded (by Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Increase db1081 weight

https://gerrit.wikimedia.org/r/346701

Change 346701 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Increase db1081 weight

https://gerrit.wikimedia.org/r/346701

Mentioned in SAL (#wikimedia-operations) [2017-04-06T07:56:25Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Increae db1081 weight - T161088 (duration: 00m 39s)

Change 346707 had a related patch set uploaded (by Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Increase db1081 weight

https://gerrit.wikimedia.org/r/346707

Change 346707 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Increase db1081 weight

https://gerrit.wikimedia.org/r/346707

Mentioned in SAL (#wikimedia-operations) [2017-04-06T08:43:39Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Increae db1081 weight - T161088 (duration: 00m 39s)

Change 346721 had a related patch set uploaded (by Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Increase db1081 weight

https://gerrit.wikimedia.org/r/346721

Change 346721 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Increase db1081 weight

https://gerrit.wikimedia.org/r/346721

Mentioned in SAL (#wikimedia-operations) [2017-04-06T11:27:08Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Increae db1081 weight - T161088 (duration: 00m 40s)

Change 346750 had a related patch set uploaded (by Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Restore db1081 original weight

https://gerrit.wikimedia.org/r/346750

Change 346750 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Restore db1081 original weight

https://gerrit.wikimedia.org/r/346750

Mentioned in SAL (#wikimedia-operations) [2017-04-06T12:28:21Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Restore db1081 original weight - T161088 (duration: 00m 40s)

Servers that need to be migrated to file per table:

db1059
db1084
db1091
db1097

root@neodymium:/home/marostegui/git/software/dbtools# cumin 'db10[68,53,56,59,64,81,84,91,97].eqiad.wmnet' 'ls -lh /srv/sqldata/ibdata1'
9 hosts will be targeted:
db[1053,1056,1059,1064,1068,1081,1084,1091,1097].eqiad.wmnet
Confirm to continue [y/n]? y
===== NODE GROUP =====
(1) db1053.eqiad.wmnet
----- OUTPUT of 'ls -lh /srv/sqldata/ibdata1' -----
-rw-rw---- 1 mysql mysql 7.8G Jul 24 08:11 /srv/sqldata/ibdata1
===== NODE GROUP =====
(1) db1081.eqiad.wmnet
----- OUTPUT of 'ls -lh /srv/sqldata/ibdata1' -----
-rw-rw---- 1 mysql mysql 1.9G Jul 24 08:11 /srv/sqldata/ibdata1
===== NODE GROUP =====
(1) db1059.eqiad.wmnet
----- OUTPUT of 'ls -lh /srv/sqldata/ibdata1' -----
-rw-rw---- 1 mysql mysql 682G Jul 24 08:11 /srv/sqldata/ibdata1
===== NODE GROUP =====
(1) db1064.eqiad.wmnet
----- OUTPUT of 'ls -lh /srv/sqldata/ibdata1' -----
-rw-rw---- 1 mysql mysql 46G Jul 24 08:11 /srv/sqldata/ibdata1
===== NODE GROUP =====
(1) db1068.eqiad.wmnet
----- OUTPUT of 'ls -lh /srv/sqldata/ibdata1' -----
-rw-rw---- 1 mysql mysql 1.6G Jul 24 08:11 /srv/sqldata/ibdata1
===== NODE GROUP =====
(1) db1097.eqiad.wmnet
----- OUTPUT of 'ls -lh /srv/sqldata/ibdata1' -----
-rw-rw---- 1 mysql mysql 1.3T Jul 24 08:11 /srv/sqldata/ibdata1
===== NODE GROUP =====
(1) db1056.eqiad.wmnet
----- OUTPUT of 'ls -lh /srv/sqldata/ibdata1' -----
-rw-rw---- 1 mysql mysql 6.0G Jul 24 08:11 /srv/sqldata/ibdata1
===== NODE GROUP =====
(2) db[1084,1091].eqiad.wmnet
----- OUTPUT of 'ls -lh /srv/sqldata/ibdata1' -----
-rw-rw---- 1 mysql mysql 1.2T Jul 24 08:11 /srv/sqldata/ibdata1
================
PASS |██████████████████████████████████████████████████████████████| 100% (9/9) [00:00<00:00, 23.02hosts/s]
FAIL |                                                                      |   0% (0/9) [00:00<?, ?hosts/s]
100.0% (9/9) success ratio (>= 100.0% threshold) for command: 'ls -lh /srv/sqldata/ibdata1'.
100.0% (9/9) success ratio (>= 100.0% threshold) of nodes successfully executed all commands
Marostegui renamed this task from Defragment s4: db1091, db1084, db1081, d1059 and probably the rest to Migrate some s4 hosts to file per table.Jul 24 2017, 8:17 AM
Marostegui triaged this task as Medium priority.
Marostegui updated the task description. (Show Details)

Change 374521 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Depool db1097

https://gerrit.wikimedia.org/r/374521

Change 374521 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Depool db1097

https://gerrit.wikimedia.org/r/374521

Mentioned in SAL (#wikimedia-operations) [2017-08-29T10:57:31Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Depool db1097 - T161088 (duration: 00m 43s)

Mentioned in SAL (#wikimedia-operations) [2017-08-29T11:08:13Z] <marostegui> Stop MariaDB on db1097 to migrate it to file per table - T161088

Change 374741 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Give more weight to db1055

https://gerrit.wikimedia.org/r/374741

Change 374741 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Give more weight to db1055

https://gerrit.wikimedia.org/r/374741

Change 387165 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/puppet@production] mariadb: Move db1103 from s3 to s4

https://gerrit.wikimedia.org/r/387165

Change 387166 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Move db1103 from s3 to s4

https://gerrit.wikimedia.org/r/387166

Change 387166 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Move db1103 from s3 to s4

https://gerrit.wikimedia.org/r/387166

Mentioned in SAL (#wikimedia-operations) [2017-10-30T06:40:53Z] <marostegui> Stop MySQL on db1103 to reclone it - T161088

Mentioned in SAL (#wikimedia-operations) [2017-10-30T06:41:22Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Move db1103 from s3 to s4 - T161088 (duration: 00m 49s)

Change 387165 merged by Marostegui:
[operations/puppet@production] mariadb: Move db1103 from s3 to s4

https://gerrit.wikimedia.org/r/387165

Change 387167 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Depool db1097

https://gerrit.wikimedia.org/r/387167

Change 387167 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Depool db1097

https://gerrit.wikimedia.org/r/387167

Mentioned in SAL (#wikimedia-operations) [2017-10-30T06:47:16Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Depool db1097 - T161088 (duration: 00m 50s)

Mentioned in SAL (#wikimedia-operations) [2017-10-30T06:51:17Z] <marostegui> Stop MySQL on db1097 to clone db1103 - T161088

Change 387169 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/software@master] s3,s4.hosts: Move db1103 from s3 to s4

https://gerrit.wikimedia.org/r/387169

Change 387169 merged by jenkins-bot:
[operations/software@master] s3,s4.hosts: Move db1103 from s3 to s4

https://gerrit.wikimedia.org/r/387169

Change 387183 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Repool db1097 with low weight

https://gerrit.wikimedia.org/r/387183

Change 387183 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Repool db1097 with low weight

https://gerrit.wikimedia.org/r/387183

Mentioned in SAL (#wikimedia-operations) [2017-10-30T08:38:09Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Repool db1097 with low weight - T161088 (duration: 00m 50s)

Change 387187 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Increase db1097 API weight

https://gerrit.wikimedia.org/r/387187

Change 387187 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Increase db1097 API weight

https://gerrit.wikimedia.org/r/387187

Mentioned in SAL (#wikimedia-operations) [2017-10-30T09:08:54Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Increase db1097 api traffic - T161088 (duration: 00m 50s)

Mentioned in SAL (#wikimedia-operations) [2017-10-30T09:57:09Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Restore db1097 original weight - T161088 (duration: 00m 52s)

Change 387522 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Depool db1084

https://gerrit.wikimedia.org/r/387522

Change 387522 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Depool db1084

https://gerrit.wikimedia.org/r/387522

Mentioned in SAL (#wikimedia-operations) [2017-10-31T06:28:53Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Depool db1084 - T161088 (duration: 00m 49s)

Mentioned in SAL (#wikimedia-operations) [2017-10-31T06:37:16Z] <marostegui> Stop MySQL on db1097 and db1084 - T161088

Change 387531 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Repool db1084 with low weight

https://gerrit.wikimedia.org/r/387531

Change 387531 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Repool db1084 with low weight

https://gerrit.wikimedia.org/r/387531

Mentioned in SAL (#wikimedia-operations) [2017-10-31T08:22:54Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Repool db1084 with low weight - T161088 (duration: 00m 50s)

Change 387533 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Increase db1084 weight

https://gerrit.wikimedia.org/r/387533

Change 387533 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Increase db1084 weight

https://gerrit.wikimedia.org/r/387533

Mentioned in SAL (#wikimedia-operations) [2017-10-31T08:46:05Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Increase db1084 weight - T161088 (duration: 00m 50s)

Change 387537 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Increase weight db1084

https://gerrit.wikimedia.org/r/387537

Change 387537 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Increase weight db1084

https://gerrit.wikimedia.org/r/387537

Mentioned in SAL (#wikimedia-operations) [2017-10-31T10:05:05Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Increase db1084 weight - T161088 (duration: 00m 50s)

Change 387545 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Restore db1084 original weight

https://gerrit.wikimedia.org/r/387545

Change 387545 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Restore db1084 original weight

https://gerrit.wikimedia.org/r/387545

Mentioned in SAL (#wikimedia-operations) [2017-10-31T10:26:17Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Restore db1084 original weight - T161088 (duration: 00m 48s)

Change 387979 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Depool db1091

https://gerrit.wikimedia.org/r/387979

Change 387979 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Depool db1091

https://gerrit.wikimedia.org/r/387979

Mentioned in SAL (#wikimedia-operations) [2017-11-02T06:30:27Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Depool db1091 - T161088 (duration: 01m 07s)

Change 387998 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Repool db1091 with low weight

https://gerrit.wikimedia.org/r/387998

Change 387998 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Repool db1091 with low weight

https://gerrit.wikimedia.org/r/387998

Mentioned in SAL (#wikimedia-operations) [2017-11-02T08:28:30Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Repool db1091 with low weight after maintenance - T161088 (duration: 00m 50s)

Change 388000 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Increase db1091 weight

https://gerrit.wikimedia.org/r/388000

Change 388000 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Increase db1091 weight

https://gerrit.wikimedia.org/r/388000

Mentioned in SAL (#wikimedia-operations) [2017-11-02T08:46:40Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Increase db1091 weight after maintenance - T161088 (duration: 00m 50s)

Change 388001 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Increase weight for db1091

https://gerrit.wikimedia.org/r/388001

Change 388001 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Increase weight for db1091

https://gerrit.wikimedia.org/r/388001

Mentioned in SAL (#wikimedia-operations) [2017-11-02T09:06:26Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Increase db1091 weight after maintenance - T161088 (duration: 00m 50s)

Change 388004 had a related patch set uploaded (by Marostegui; owner: Marostegui):
[operations/mediawiki-config@master] db-eqiad.php: Restore db1091 original weight

https://gerrit.wikimedia.org/r/388004

Change 388004 merged by jenkins-bot:
[operations/mediawiki-config@master] db-eqiad.php: Restore db1091 original weight

https://gerrit.wikimedia.org/r/388004

Mentioned in SAL (#wikimedia-operations) [2017-11-02T09:41:49Z] <marostegui@tin> Synchronized wmf-config/db-eqiad.php: Restore db1091 original weight - T161088 (duration: 00m 50s)

db1091 is done and its original traffic weight is restored.