====== Решение Ceph ======
* [[https://ru.wikipedia.org/wiki/Ceph|Ceph — свободная программная объектная сеть хранения (Материал из Википедии)]]
* [[https://ceph.io/get/|Официальный сайт объектной системы хранения Ceph/GETTING STARTED]]
* [[https://habr.com/ru/post/313644/|Знакомство с хранилищем Ceph в картинках (Хабр)]]
* [[https://habr.com/ru/company/oleg-bunin/blog/431536/|Ceph. Анатомия катастрофы (Хабр)]]
* [[http://onreader.mdl.ru/LearningCeph/content/index.html|Практическое руководство по разработке, реализации и управлению вашей управляемой программным обеспечением, массивно масштабируемой системы хранения Ceph]]
* [[http://onreader.mdl.ru/MasteringCeph.2ed/content/index.html|Полное руководство Ceph, 2е изд.]]
* [[http://onreader.mdl.ru/CephCookbook/content/index.html|Книга рецептов Ceph]]
===== Развертывание =====
* [[https://docs.ceph.com/en/latest/install/manual-deployment/|MANUAL DEPLOYMENT]]
* [[https://www.server-world.info/en/note?os=Debian_11&p=ceph14&f=1|Debian 11 : Ceph Nautilus]]
==== Установка ====
nodeN# apt install ceph
==== MON ====
* [[https://docs.ceph.com/en/latest/man/8/ceph-mon/|CEPH-MON – CEPH MONITOR DAEMON]]
=== Начальная конфигурация ===
nodeN# cat /etc/ceph/ceph.conf
[global]
fsid = a7f64266-0894-4f1e-a635-d0aeaca0e993
mon initial members = node3,node4,node5
mon host = 192.168.X.3,192.168.X.4,192.168.X.5
public network = 192.168.X.0/24
auth cluster required = none
auth service required = none
auth client required = none
scp /etc/ceph/ceph.conf node4:/etc/ceph/
scp /etc/ceph/ceph.conf node5:/etc/ceph/
monmaptool --create --add node3 192.168.X.3 --fsid a7f64266-0894-4f1e-a635-d0aeaca0e993 /tmp/monmap
monmaptool --add node4 192.168.X.4 --fsid a7f64266-0894-4f1e-a635-d0aeaca0e993 /tmp/monmap
monmaptool --add node5 192.168.X.5 --fsid a7f64266-0894-4f1e-a635-d0aeaca0e993 /tmp/monmap
monmaptool --print /tmp/monmap
scp /tmp/monmap node4:/tmp/
scp /tmp/monmap node5:/tmp/
node3# bash -c '
sudo -u ceph mkdir /var/lib/ceph/mon/ceph-node3
ssh node4 sudo -u ceph mkdir /var/lib/ceph/mon/ceph-node4
ssh node5 sudo -u ceph mkdir /var/lib/ceph/mon/ceph-node5
'
node3# bash -c '
sudo -u ceph ceph-mon --mkfs -i node3 --monmap /tmp/monmap
ssh node4 sudo -u ceph ceph-mon --mkfs -i node4 --monmap /tmp/monmap
ssh node5 sudo -u ceph ceph-mon --mkfs -i node5 --monmap /tmp/monmap
'
node3# ls -l /var/lib/ceph/mon/ceph-node3
node3# bash -c '
systemctl start ceph-mon@node3
systemctl enable ceph-mon@node3
ssh node4 systemctl start ceph-mon@node4
ssh node4 systemctl enable ceph-mon@node4
ssh node5 systemctl start ceph-mon@node5
ssh node5 systemctl enable ceph-mon@node5
'
* [[https://www.suse.com/support/kb/doc/?id=000019960|Cluster status shows a "mons are allowing insecure global_id reclaim" health warning]]
debian11_12/ubuntu# ceph mon enable-msgr2
debian11_12/ubuntu# ceph config set mon mon_warn_on_insecure_global_id_reclaim false
debian11_12/ubuntu# ceph config set mon mon_warn_on_insecure_global_id_reclaim_allowed false
# ceph -s
# ps ax | grep ceph
=== Поиск проблемных узлов ===
# ceph health detail
MON_DOWN 1/3 mons down, quorum node4,node5
mon.node3 (rank 0) addr 192.168.X.3:6789/0 is down (out of quorum)
==== MGR ====
* [[https://docs.ceph.com/en/latest/mgr/index.html|CEPH MANAGER DAEMON]]
node3# bash -c '
sudo -u ceph mkdir /var/lib/ceph/mgr/ceph-node3
systemctl start ceph-mgr@node3
systemctl enable ceph-mgr@node3
ssh node4 sudo -u ceph mkdir /var/lib/ceph/mgr/ceph-node4
ssh node4 systemctl start ceph-mgr@node4
ssh node4 systemctl enable ceph-mgr@node4
ssh node5 sudo -u ceph mkdir /var/lib/ceph/mgr/ceph-node5
ssh node5 systemctl start ceph-mgr@node5
ssh node5 systemctl enable ceph-mgr@node5
'
# ceph -s
# ps ax | grep ceph
==== OSD ====
* [[https://docs.ceph.com/en/latest/man/8/ceph-osd/|CEPH-OSD – CEPH OBJECT STORAGE DAEMON]]
=== Начальная конфигурация ===
# lsblk
node3# bash -c '
ceph-volume lvm create --data /dev/sdb
ssh node4 ceph-volume lvm create --data /dev/sdb
ssh node5 ceph-volume lvm create --data /dev/sdb
'
# vgs
# ps ax | grep ceph
# ceph -s
# ceph osd stat
# ceph osd tree
# ceph osd df
=== Добавление узла OSD ===
node3# scp /etc/ceph/ceph.conf node6:/etc/ceph/
node3# ssh node6 ceph-volume lvm create --data /dev/sdb
node3# watch ceph -s
=== Удаление узла OSD ===
ceph osd safe-to-destroy osd.0
OSD(s) 0 are safe to destroy without reducing data durability.
ceph osd destroy 0 --yes-i-really-mean-it
ceph osd purge 0 --yes-i-really-mean-it
=== Управление приоритетом использования узла OSD ===
ceph osd crush reweight osd.2 0
ceph osd crush reweight osd.2 0.00389
node3# ceph osd df
ID CLASS WEIGHT REWEIGHT SIZE USE AVAIL %USE VAR PGS
0 hdd 0 0.83000 4.00GiB 1.01GiB 2.99GiB 25.30 0.37 0
1 hdd 0.00389 0.83000 4.00GiB 2.78GiB 1.21GiB 69.66 1.02 74
2 hdd 0.00389 0.83000 4.00GiB 3.04GiB 980MiB 76.04 1.11 81
3 hdd 0.00389 1.00000 4.00GiB 3.43GiB 575MiB 85.96 1.26 99
4 hdd 0.00389 1.00000 4.00GiB 3.38GiB 631MiB 84.58 1.24 102
ceph osd reweight osd.2 1
==== OSD POOL ====
* [[https://docs.ceph.com/en/latest/rados/operations/pools/|POOLS]]
=== Расчет параметров и создание POOL ===
* [[https://docs.ceph.com/en/latest/rados/operations/placement-groups/|PLACEMENT GROUPS]]
Total PGs = OSDs * 100 / pool_size
5*100/3 = 166,6...
# ceph osd pool create test-pool1 180
debian11_12/ubuntu# ceph osd pool create test-pool1 128
или
debian11_12/ubuntu# ceph config set global mon_warn_on_pool_pg_num_not_power_of_two false
# ceph -s
...
pgs: 180 active+clean
# ceph pg dump
# ceph osd lspools
=== Изменение параметров POOL ===
ceph osd pool set test-pool1 pg_num 250
ceph osd pool set test-pool1 pgp_num 250
ceph osd pool get test-pool1 size
ceph osd pool get test-pool1 min_size
ceph osd pool set test-pool1 size 2
ceph osd pool set test-pool1 size 3
==== RBD POOL ====
* Reliable Autonomic Distributed Object Store = RADOS
* [[https://docs.ceph.com/en/latest/rbd/rados-rbd-cmds/|BASIC BLOCK DEVICE COMMANDS]]
* [[https://docs.ceph.com/en/latest/man/8/rbd/|rbd – manage rados block device (RBD) images]]
# ceph df
# rbd pool init test-pool1
# rbd create -p test-pool1 rbd1 --size 1G
# rbd list test-pool1
# rbd info test-pool1/rbd1
# rbd resize --size 3584M test-pool1/rbd1
node3# bash -c '
systemctl restart tgt
ssh node4 systemctl restart tgt
ssh node5 systemctl restart tgt
'
==== TGT-RBD ====
* [[https://habr.com/ru/post/315646/|Кластер хранилища Ceph на VMWare за 10 минут]]
nodeN# apt install tgt-rbd
nodeN# cat /etc/tgt/conf.d/ceph.conf
driver iscsi
bs-type rbd
backing-store test-pool1/rbd1
initiator-address ALL
scp /etc/tgt/conf.d/ceph.conf node4:/etc/tgt/conf.d/
scp /etc/tgt/conf.d/ceph.conf node5:/etc/tgt/conf.d/
node3# bash -c '
systemctl restart tgt
ssh node4 systemctl restart tgt
ssh node5 systemctl restart tgt
'
systemctl status tgt
===
https://ceph.io/planet/find-the-osd-location/
# ceph osd find 0
ceph health detail
===
===
https://medium.com/flant-com/3-cases-from-our-ceph-experience-c46efea0e527
Case #1. Gracefully taking out OSDs from the Ceph cluster
...
ceph tell 'osd.*' injectargs --osd-max-backfills=1
...