Commit Graph

503 Commits

Author SHA1 Message Date
Jakub Sokołowski f96fe517e7
beacon-node-builds: allow staff group membets to start builds
I was thinking of defining the jobs as nimbus user systemd units, but
that would not work unless nimbus user logged in at least once.

And Systemd services have no functionality for allowing non-root users
to start/stop/restart system services, so sudo is the simplest way.

Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-02-08 12:36:02 +01:00
Jakub Sokołowski e5037a6751
geth: bump instance data volume to 1200 GB
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-02-06 20:38:33 +01:00
Jakub Sokołowski 8e0966fb0e
beacon-node-builds: don't fetch if we're on a detached head
Causes failures when a tag is checked out, nothing to fetch.

Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-02-04 23:48:26 +01:00
zah 33ba3b1e8b Upgrade the production fleet to 1.0.7 2021-02-04 21:41:22 +01:00
Jakub Sokołowski b7226818d3
disable kibana-logtrail, it's not been updated to 7.10.2
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-02-01 16:05:50 +01:00
Jakub Sokołowski dbac764185
upgrade ElasticSearch to 7.10.2
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-02-01 15:57:10 +01:00
Jakub Sokołowski 921b1c3ea7
ansible.cfg: enable ansible ssh multiplexing
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-02-01 13:21:44 +01:00
Jakub Sokołowski 157d215230
specify beacon_node_netkey for nimbus-mainnet-small bootnodes
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-29 13:26:55 +01:00
Jakub Sokołowski 87a0503c22
increase volume size for Mainnet Geth to 1TB
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-27 16:47:25 +01:00
Ștefan Talpalaru 37b4e558b2 fix the Makefile and a shell script
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-26 15:31:26 +01:00
Jakub Sokołowski 26a31c5d63
use 8000+ UIDs for extra SSH users to not clash with core
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-26 12:11:33 +01:00
Jakub Sokołowski d12ccc626b
ansible.cfg: shell instead of command as default module
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-25 15:10:25 +01:00
Jakub Sokołowski 8f76c6810a
log-store: fix mounting of data volumes
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-25 15:08:34 +01:00
Jakub Sokołowski 62f44b8a05
drop old fleet from group and host vars and pyrmont playbook
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-22 15:59:47 +01:00
Jakub Sokołowski 49bd1fa267
remove old pyrmont fleet, move new one back to pyrmont.tf
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-22 15:56:26 +01:00
Jakub Sokołowski 8aa381ca2e
update names in inventory for nimbus.mainnet hosts
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-22 15:46:04 +01:00
Jakub Sokołowski 84be761116
fix hostname in nimbus-pyrmont-unstable-libp2p validators layout
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-22 15:19:58 +01:00
Jakub Sokołowski 8920b57c34 rename mainnet hosts to match stable branch
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-22 14:27:39 +01:00
Jakub Sokołowski 37c1c62a5d add validators distributio for the new fleet
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-22 14:27:39 +01:00
Jakub Sokołowski 8ad0027936
log-store: add 2GB swap files to hosts
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-20 20:35:33 +01:00
Jakub Sokołowski 1218ab1772
lower volume size for logs to 200 GB for now
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-20 20:35:16 +01:00
Jakub Sokołowski 9a9619138b
add dashboard link to README
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-20 19:27:04 +01:00
Jakub Sokołowski 6e138997e2
re-enable sending logs to logstash
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-20 19:27:04 +01:00
Jakub Sokołowski 45a0315bb0
deploy elasticsearch and a kibana dashboard
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-20 16:00:29 +01:00
Jakub Sokołowski ef8e77a1ac
re-add the ELK stack hosts
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-20 15:59:11 +01:00
Jakub Sokołowski 8d939e1cd9
bump data volume IOPS to 2500
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-19 16:29:53 +01:00
Jakub Sokołowski 2b4c2320ce
move pyrmont variables to pyrmont.tf
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-19 10:33:27 +01:00
Jakub Sokołowski 67ebb93be3
fix wrong build nodes listed in playbook
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-18 19:53:52 +01:00
Jakub Sokołowski 0bc722ca21
dist-validators: don't run any tasks if no layout is defined
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-18 19:47:51 +01:00
Jakub Sokołowski 89dfe7ca5c
add new fleet hosts to inventory
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-18 19:35:03 +01:00
Jakub Sokołowski 2b44636cd6
create new pyrmont fleet using new branch naming
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-18 19:04:10 +01:00
Jakub Sokołowski 05684f23b1
drop unused toledor and maste references
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-18 16:52:40 +01:00
Jakub Sokołowski e2a2070f43
bump data volume IOPS to 2000
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-18 16:14:02 +01:00
Jakub Sokołowski 2784affaa9
pyrmont: make data EBS volumes to io1 type, 1500 IOPS
The max IOPS for gp2/3 is 16,000 for 16 TB.
The max ratio of IOPS to size is 50.
This makes the max of 7500 IOPS for 150 GB.

Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-18 15:51:12 +01:00
Zahary Karadjov e2244166ca
Simple Makefile target for generating a SSH config file with the fleet hosts
`make ssh-config` will generate a file ~/.ssh/config.d/infra-numbus with all
hostnames from the Ansible inventory. You can `Include` this file from your
main config file and re-generate it at will at any time. Having such a config
file makes it easier to ssh into any host on the fleet by taking advantage of
the built-in ssh hostname autocompletion in your shell.

The names of the entries were chosen in a way such that typing

ssh nimbus<tab>goerli<tab>

... produces the expected result.
2021-01-18 15:09:08 +01:00
Jakub Sokołowski ac608401ae
pyrmont: use gp3 EBS volumes for data on devel-small fleet
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-14 17:14:21 +01:00
Jakub Sokołowski 2132851b16
use bootstrap__active_extra_users variable
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-14 14:01:57 +01:00
Jakub Sokołowski 701ca85795
geth: lower size of goerli host to t3a.medium
More memory is only necessary when syncing from scratch.

Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-14 12:08:42 +01:00
Jakub Sokołowski ab017c29bb
update inventory
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-13 17:07:31 +01:00
Jakub Sokołowski 7660970894
bump size of geth mainnet node to t3a.xlarge temporarily
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2021-01-13 17:07:20 +01:00
Ștefan Talpalaru e71d1643b8 beacon-node-builds: support core dumps
- build with "-d:noSignalHandler" in NIMFLAGS
- keep binary copies for a week, so we can use them with core dumps on
  target hosts
- finish replacing "item.branch" with "item.version"
- manually add a [nimbus-master] section to "ansible/inventory/test"
2020-12-18 21:56:32 +02:00
Jakub Sokołowski 8dc1db8f0b
nimbus-builds: add release build for eth2.prod fleet
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2020-12-11 18:44:34 +01:00
Jakub Sokołowski 6ecba424d1
drop useless debug task
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2020-12-04 18:27:17 +01:00
Jakub Sokołowski 48529cb841
use Infura for Web3 on master-large-01 on Pyrmont for comparison
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2020-12-04 18:26:05 +01:00
Jakub Sokołowski 9c5abd392a
allow VPN connections to Goerli/Mainnet WebSocket API
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2020-12-04 18:20:05 +01:00
Jakub Sokołowski 989aeacb51
use Geth Goerli WebSocket endpoint for Pyrmont fleet
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2020-12-04 13:09:48 +01:00
Jakub Sokołowski 65f0d944d8
bump max logs for logrotate to 72
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2020-11-26 21:30:32 +01:00
Jakub Sokołowski 996e2df2e8
make docker log files contain only container messages
See: https://github.com/status-im/infra-nimbus/issues/26

Signed-off-by: Jakub Sokołowski <jakub@status.im>
2020-11-26 19:06:08 +01:00
Jakub Sokołowski 796c7ce292
bump count of logrotate files to keep to 24
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2020-11-26 18:30:05 +01:00
Jakub Sokołowski 176d843a40
drop toledo hosts
Signed-off-by: Jakub Sokołowski <jakub@status.im>
2020-11-26 17:06:35 +01:00