Cephadm est capable de mettre à niveau et en toute sécurité un cluster Ceph d'une version de correction de bogue à une autre. Par exemple, vous pouvez passer de la v15.2.0 (la première version d'Octopus) à la prochaine version ponctuelle v15.2.1. Vous pouvez aussi passer a une version majeur de la même maniere ;)
Le processus de mise à niveau automatisé suit les meilleures pratiques Ceph. Par exemple:
- L'ordre de mise à niveau commence avec les gestionnaires, les moniteurs, puis les autres démons.
- Chaque démon est redémarré uniquement après que Ceph indique que le cluster restera disponible.
- Gardez à l'esprit que l'état d'intégrité du cluster Ceph est susceptible de basculer sur HEALTH_WARNING pendant la mise à niveau.
mise à jour via la commande ceph orch upgrade vers la version 16.2.0
# vérification de l'etat du cluster
[ceph: root@cn1 /]# ceph -s
cluster:
id: be024bba-5f15-11eb-ad27-5254001065ff
health: HEALTH_OK
services:
mon: 5 daemons, quorum cn1,cn2,cn3,cnrgw1,cn4 (age 18m)
mgr: cn1.vvraig(active, since 18m), standbys: cn2.zphxce
osd: 8 osds: 8 up (since 18m), 8 in (since 3w)
rgw: 2 daemons active (demodom.fr-est-1.cnrgw1.umtgun, demodom.fr-est-2.cnrgw2.fggict)
task status:
data:
pools: 13 pools, 289 pgs
objects: 1.28k objects, 29 KiB
usage: 8.8 GiB used, 351 GiB / 360 GiB avail
pgs: 289 active+clean
io:
client: 11 KiB/s rd, 0 B/s wr, 10 op/s rd, 10 op/s wr
progress:
Upgrade to docker.io/ceph/ceph:v16.2.0 (0s)
[............................]
# dans quelle version je suis
[ceph: root@cn1 /]# ceph versions
{
"mon": {
"ceph version 15.2.9 (357616cbf726abb779ca75a551e8d02568e15b17) octopus (stable)": 5
},
"mgr": {
"ceph version 15.2.9 (357616cbf726abb779ca75a551e8d02568e15b17) octopus (stable)": 2
},
"osd": {
"ceph version 15.2.9 (357616cbf726abb779ca75a551e8d02568e15b17) octopus (stable)": 8
},
"mds": {},
"rgw": {
"ceph version 15.2.9 (357616cbf726abb779ca75a551e8d02568e15b17) octopus (stable)": 2
},
"overall": {
"ceph version 15.2.9 (357616cbf726abb779ca75a551e8d02568e15b17) octopus (stable)": 17
}
}
# migration avec une commande
[ceph: root@cn1 /]# ceph orch upgrade start --ceph-version 16.2.0
Initiating upgrade to docker.io/ceph/ceph:v16.2.0
# vérification de ce qui se passe
[ceph: root@cn1 /]# ceph versions
{
"mon": {
"ceph version 15.2.9 (357616cbf726abb779ca75a551e8d02568e15b17) octopus (stable)": 3,
"ceph version 16.2.0 (0c2054e95bcd9b30fdd908a79ac1d8bbc3394442) pacific (stable)": 2
},
"mgr": {
"ceph version 16.2.0 (0c2054e95bcd9b30fdd908a79ac1d8bbc3394442) pacific (stable)": 2
},
"osd": {
"ceph version 15.2.9 (357616cbf726abb779ca75a551e8d02568e15b17) octopus (stable)": 8
},
"mds": {},
"rgw": {
"ceph version 15.2.9 (357616cbf726abb779ca75a551e8d02568e15b17) octopus (stable)": 2
},
"overall": {
"ceph version 15.2.9 (357616cbf726abb779ca75a551e8d02568e15b17) octopus (stable)": 13,
"ceph version 16.2.0 (0c2054e95bcd9b30fdd908a79ac1d8bbc3394442) pacific (stable)": 4
}
}
# Remarque: cela commence par les mgr, puis les mon
[ceph: root@cn1 /]# ceph -s
cluster:
id: be024bba-5f15-11eb-ad27-5254001065ff
health: HEALTH_OK
services:
mon: 5 daemons, quorum cn1,cn2,cn3,cnrgw1,cn4 (age 66s)
mgr: cn1.vvraig(active, since 99s), standbys: cn2.zphxce
osd: 8 osds: 8 up (since 26m), 8 in (since 3w)
rgw: 2 daemons active (2 hosts, 2 zones)
data:
pools: 13 pools, 289 pgs
objects: 1.39k objects, 29 KiB
usage: 8.8 GiB used, 351 GiB / 360 GiB avail
pgs: 289 active+clean
io:
client: 511 B/s rd, 0 op/s rd, 0 op/s wr
progress:
Upgrade to 16.2.0 (16s)
[===.........................] (remaining: 112s)
# le cluster est toujours oppérationnel
[ceph: root@cn1 /]# ceph -W cephadm
cluster:
id: be024bba-5f15-11eb-ad27-5254001065ff
health: HEALTH_OK
services:
mon: 5 daemons, quorum cn1,cn2,cn3,cnrgw1,cn4 (age 5m)
mgr: cn1.vvraig(active, since 11m), standbys: cn2.zphxce
osd: 8 osds: 8 up (since 13s), 8 in (since 3w)
rgw: 1 daemon active (1 hosts, 1 zones)
data:
pools: 13 pools, 289 pgs
objects: 1.39k objects, 29 KiB
usage: 1.2 GiB used, 359 GiB / 360 GiB avail
pgs: 289 active+clean
io:
client: 33 KiB/s rd, 0 B/s wr, 33 op/s rd, 14 op/s wr
progress:
Upgrade to 16.2.0 (10m)
[===================.........] (remaining: 4m)
2021-04-04T01:26:35.817499+0000 mgr.cn1.vvraig [INF] Upgrade: Setting container_image for all rgw
2021-04-04T01:26:36.132912+0000 mgr.cn1.vvraig [INF] Upgrade: Setting container_image for all rbd-mirror
2021-04-04T01:26:36.201017+0000 mgr.cn1.vvraig [INF] Upgrade: Setting container_image for all iscsi
2021-04-04T01:26:36.280148+0000 mgr.cn1.vvraig [INF] Upgrade: Setting container_image for all nfs
2021-04-04T01:26:38.776443+0000 mgr.cn1.vvraig [INF] Upgrade: Updating node-exporter.cn1
2021-04-04T01:26:38.777344+0000 mgr.cn1.vvraig [INF] Deploying daemon node-exporter.cn1 on cn1
2021-04-04T01:26:47.806717+0000 mgr.cn1.vvraig [INF] Upgrade: Updating node-exporter.cn2
2021-04-04T01:26:47.806961+0000 mgr.cn1.vvraig [INF] Deploying daemon node-exporter.cn2 on cn2
2021-04-04T01:26:54.899946+0000 mgr.cn1.vvraig [INF] Upgrade: Updating node-exporter.cn3
2021-04-04T01:26:54.900154+0000 mgr.cn1.vvraig [INF] Deploying daemon node-exporter.cn3 on cn3
2021-04-04T01:27:01.640376+0000 mgr.cn1.vvraig [INF] Upgrade: Updating node-exporter.cn4
2021-04-04T01:27:01.640554+0000 mgr.cn1.vvraig [INF] Deploying daemon node-exporter.cn4 on cn4
2021-04-04T01:27:06.941037+0000 mgr.cn1.vvraig [INF] Upgrade: Updating node-exporter.cnrgw1
2021-04-04T01:27:06.941244+0000 mgr.cn1.vvraig [INF] Deploying daemon node-exporter.cnrgw1 on cnrgw1
2021-04-04T01:27:11.666549+0000 mgr.cn1.vvraig [INF] Upgrade: Updating node-exporter.cnrgw2
2021-04-04T01:27:11.666826+0000 mgr.cn1.vvraig [INF] Deploying daemon node-exporter.cnrgw2 on cnrgw2
2021-04-04T01:27:16.813336+0000 mgr.cn1.vvraig [INF] Upgrade: Updating prometheus.cn1
2021-04-04T01:27:16.886056+0000 mgr.cn1.vvraig [INF] Deploying daemon prometheus.cn1 on cn1
2021-04-04T01:27:27.235493+0000 mgr.cn1.vvraig [INF] Upgrade: Updating alertmanager.cn1
2021-04-04T01:27:27.248582+0000 mgr.cn1.vvraig [INF] Deploying daemon alertmanager.cn1 on cn1
2021-04-04T01:27:40.359685+0000 mgr.cn1.vvraig [INF] Upgrade: Updating grafana.cn1
2021-04-04T01:27:40.783075+0000 mgr.cn1.vvraig [INF] Deploying daemon grafana.cn1 on cn1
2021-04-04T01:28:50.176234+0000 mgr.cn1.vvraig [INF] Upgrade: Finalizing container_image settings
2021-04-04T01:28:50.739604+0000 mgr.cn1.vvraig [INF] Upgrade: Complete!
# remarque: c'est plus simple a suivre
[ceph: root@cn1 /]# ceph versions
{
"mon": {
"ceph version 16.2.0 (0c2054e95bcd9b30fdd908a79ac1d8bbc3394442) pacific (stable)": 5
},
"mgr": {
"ceph version 16.2.0 (0c2054e95bcd9b30fdd908a79ac1d8bbc3394442) pacific (stable)": 2
},
"osd": {
"ceph version 16.2.0 (0c2054e95bcd9b30fdd908a79ac1d8bbc3394442) pacific (stable)": 8
},
"mds": {},
"rgw": {
"ceph version 16.2.0 (0c2054e95bcd9b30fdd908a79ac1d8bbc3394442) pacific (stable)": 2
},
"overall": {
"ceph version 16.2.0 (0c2054e95bcd9b30fdd908a79ac1d8bbc3394442) pacific (stable)": 17
}
}
# c'est migré en 16.2.0
[ceph: root@cn1 /]# ceph -s
cluster:
id: be024bba-5f15-11eb-ad27-5254001065ff
health: HEALTH_OK
services:
mon: 5 daemons, quorum cn1,cn2,cn3,cnrgw1,cn4 (age 8m)
mgr: cn1.vvraig(active, since 14m), standbys: cn2.zphxce
osd: 8 osds: 8 up (since 3m), 8 in (since 3w)
rgw: 2 daemons active (2 hosts, 2 zones)
data:
pools: 13 pools, 289 pgs
objects: 1.39k objects, 29 KiB
usage: 1.2 GiB used, 359 GiB / 360 GiB avail
pgs: 289 active+clean
io:
client: 27 KiB/s rd, 0 B/s wr, 27 op/s rd, 15 op/s wr
# ok tout fonctionne
[ceph: root@cn1 /]#