Commit Graph

214 Commits

Author SHA1 Message Date
Joshua Boniface 0aa328e350 Add PostgreSQL to daily backup script 2021-06-01 10:10:22 -04:00
Joshua Boniface 9deee94332 Update tags and fix backup keys to var 2021-05-27 12:29:19 -04:00
Joshua Boniface cae8cfc4cb Add norestart policy for apt updates 2021-05-27 01:38:43 -04:00
Joshua Boniface 491ea77306 Add README and daemon upgrade playbook, cleanups 2021-05-20 11:02:47 -04:00
Joshua Boniface e76832de91 Allow inter-cluster orphan NTP sync
Due to the requirement of Ceph to have all peer nodes tightly
synchronized with each other to come online, PVC nodes need a way to
synchronize to each other even in the absence of an external time
reference. This is especially prevalent if a set of nodes are left
offline for an extended period (>1-2 weeks), since their hardware clocks
will drift. If the resulting Internet connectivity is then dependent on
a VM, this will cause a catch-22 and the cluster will not properly
start.

This configuration will accomplish that - if no suitable >6 stratum
peers are found, the hosts will enter orphan mode. Since they are now
all configured as "peers" with each other, they will collectively decide
on one of them to become the source and sync to it. A local stratum 10
fudge is added so that at least one of the nodes can become this source.

While this is not an ideal use of NTP, it is by far the cleanest
solution to this problem, and does not impact normal functionality when
the two configured stratum-2 servers are reachable.
2021-05-19 11:03:18 -04:00
Joshua Boniface 238449904f Move some other tasks to bootstrap role
Avoids an issue where the pvcnoded service is stopped on non-bootstrap
runs.
2021-05-13 10:17:38 -04:00
Joshua Boniface 7536732f30 Remove GRUB config from base role
This is not actually ideal.
2021-05-12 14:55:57 -04:00
Joshua Boniface 04bc9730a0 Fix version sorting bugs in kernel-cleanup.sh 2021-05-12 14:40:18 -04:00
Joshua Boniface 45322e0f9e Add additional items to base role
Backups, GRUB configuration, and IPMI configuration.
2021-05-12 13:53:15 -04:00
Joshua Boniface da9eafcdfa Fix sudoers to use conditional deploy_username 2021-04-13 16:50:05 -04:00
Joshua Boniface 70ba4b240f Allow configurable fail2ban IPs 2021-04-13 16:44:49 -04:00
Joshua Boniface ce3554b530 Allow customization of deploy username 2021-04-13 11:30:42 -04:00
Joshua Boniface 593a81e07c Fix group_vars to match new setup 2021-04-08 14:15:11 -04:00
Joshua Boniface 3819cd87fd Move to more dynamic apt configs
Allow specifying repository URLs in the group_vars, and add
release-specific template files to support future version changes.
2021-04-08 14:14:25 -04:00
Joshua Boniface 3e1d3a90b0 Update root password in default group_vars 2021-04-08 14:08:21 -04:00
Joshua Boniface 404751f695 Update relative path to bootstrap files 2021-04-08 14:04:56 -04:00
Joshua Boniface 622cef1586 Remove superfluous symlink 2021-04-08 13:50:47 -04:00
Joshua Boniface 6589a9cd38 Add sensible sorting of kernel removals 2021-04-08 13:46:43 -04:00
Joshua Boniface 6598637e91 Remove cruft and add mkpasswd setup 2021-04-08 13:46:30 -04:00
Joshua Boniface 25674731cd Update file copyright header 2021-03-25 16:58:58 -04:00
Joshua Boniface dcd0b48d94 Correct bad indentation in base role 2021-03-18 09:36:49 -04:00
Joshua Boniface 82fa85834a Add libguestfs-tools to libvirt role deps 2021-03-15 13:39:37 -04:00
Joshua Boniface 510db0df58 Add cleanup to update oneshot playbook 2021-02-02 15:41:38 -05:00
Joshua Boniface ca3a5e144f Update tags and add kernel-cleanup script 2021-02-02 15:41:38 -05:00
Joshua Boniface 1c05c8729f Fix incorrect systemd enabling in Patroni 2021-01-28 16:28:02 -05:00
Joshua Boniface 4b179b66ed Add reboot to purge 2021-01-28 14:13:15 -05:00
Joshua Boniface 71edb9db15 Remove log dirs during purge 2021-01-28 14:12:40 -05:00
Joshua Boniface f4974d648d Add some additional compression libraries 2021-01-28 13:34:58 -05:00
Joshua Boniface fa0aeec88e Add local domain to resolver config 2021-01-28 13:34:26 -05:00
Joshua Boniface 04ca8f73d2 Correct bugs during bootstrap
1. Ensure Zookeeper restarts and checks out successfully before
proceeding with other steps.
2. Make sure PVC itself doesn't start prematurely.
2021-01-28 13:32:36 -05:00
Joshua Boniface 21e3e0e172 Add removal of Zookeeper keys too 2021-01-28 13:26:46 -05:00
Joshua Boniface 20d802f0b0 Update purge script 2021-01-27 17:08:38 -05:00
Joshua Boniface b7f251ea16 Retry pgsql bootstrap startup 6 times
This will sometimes fail, so retry it several times
2021-01-27 15:45:36 -05:00
Joshua Boniface 7b08610efa Retry msgr2 enabling 6 times
This will sometimes fail, so retry it several times
2021-01-27 14:13:09 -05:00
Joshua Boniface c4c285c7b3 Remove invalid timezone entries in postgres conf 2021-01-26 15:20:25 -05:00
Joshua Boniface 97869ca5c3 Reorder Ceph stop and lower some waits 2021-01-07 11:11:16 -05:00
Joshua Boniface d35250b870 Add tasks to verify node has finished (un)flushing 2021-01-07 10:49:23 -05:00
Joshua Boniface cd164d1984 Increase all wait timeouts to 30s
Ensure that even on slow(er) clusters, these timeouts have more time to
complete before proceeding so the task won't fail.
2021-01-05 16:17:19 -05:00
Joshua Boniface 7585553225 Add default values 2020-12-21 00:20:45 -05:00
Joshua Boniface ac071f4bf0 Add configurable ZK memory limits 2020-12-21 00:20:45 -05:00
Joshua Boniface 98e3e39570 Remove libjemalloc package 2020-12-21 00:20:45 -05:00
Joshua Boniface 8e104113d7 Tune Zookeeper memory usage
Use Xms and Xmx=128M to reduce overall Zookeeper memory usage.
2020-12-21 00:20:45 -05:00
Joshua Boniface de04105a38 Add tuning for Ceph OSDs 2020-12-21 00:20:45 -05:00
Joshua Boniface 28c86d170f Don't use libjemalloc for Ceph daemons
This was an artifact of a much, much older Ceph configuration I ran, and
is not relevant with newer Ceph versions like those used in PVC.
Performance testing with Nautilus and Bluestore reveals a minimal
performance hit, and using `jemalloc` prevents cache autotuning from
being effective, so remove it.
2020-12-21 00:20:45 -05:00
Joshua Boniface f277acc974 Disable pvc-flush service while rebooting
Prevents the flush daemon from starting on node boot, before the
playbook is actually ready to unflush the node.
2020-12-15 14:32:50 -05:00
Joshua Boniface 8b474760ed Tweak oneshot script
Cleanly stop daemons; check if OSDs are back before continuing; wait
less
2020-11-26 10:51:54 -05:00
Joshua Boniface cb96ef4e7a Use new init command location
Command was renamed in the PVC CLI to facilitate other "task" actions
like backup/restore.
2020-11-24 12:22:34 -05:00
Joshua Boniface 3c0c3e8e56 Add jute.maxbuffer to Zookeeper environment ops
Adds this option based on the findings of
https://github.com/python-zk/kazoo/issues/630, whereby restores of >1MB
in size would fail. This is considered an unsafe option, but given our
usecase no actual znode should ever exceed this limit; this is purely
for the large transactions that come from a `pvc task restore` action to
an empty Zookeeper instance.
2020-11-24 12:20:25 -05:00
Joshua Boniface da8c357d38 Add PVC status MOTD script 2020-11-17 12:48:53 -05:00
Joshua Boniface 9f84609808 Set proper mode on agent plugins 2020-10-27 15:48:57 -04:00