#ceph — Public Fediverse posts

🖥️ Proxmox VE extinde oficial suportul dincolo de x86, adăugând arhitectura Arm64! Lansarea oficială a ediției Proxmox Virtual Environment (VE) 9.2 pentru Arm64 aduce suita completă de virtualizare pe servere bazate pe arhitectură ARM. Proiectul împărtășește același depozit de cod, funcționalități și ciclu de viață cu versiunea x86-64. ✨ Principalele noutăți și specificații tehnice:🚀 Paritate completă de funcționalități:• Oferă suport nativ pentru KVM, containere LXC, sistemele de fișiere/stocare ZFS și Ceph, clustering HA, backup și interfața web clasică de administrare. 🖥️ Hardware certificat & Suport Enterprise:• Optimizat pentru servere AI & Cloud: Certificare oficială din prima zi pentru platformele de servere NVIDIA Grace Hopper™ și NVIDIA Vera™.• Cerințe hardware strict server: Necesită boot prin UEFI și descrierea componentelor prin ACPI (compatibil cu ARMv9-A și ARMv8-A în regim best-effort). ⚠️ Fără plăci de dezvoltare (SBC):• Dispozitivele bazate exclusiv pe Device Tree (cum ar fi Raspberry Pi, Orange Pi etc.) NU sunt suportate oficial.⚙️ Stack tehnologic la bază:• Construiți pe Debian 13.5 "Trixie", kernel Linux 7.0, QEMU 11.0, LXC 7.0 și ZFS 2.4. 🔄 Diferențe specifice arhitecturii:• Mașinile virtuale pe Arm64 folosesc exclusiv AAVMF (UEFI) (SeaBIOS nu este disponibil).• Migrarea live a VM-urilor se poate face doar între noduri din aceeași arhitectură (x86 la x86, ARM la ARM). Un pas strategic masiv pentru datacenter-e și mediile enterprise care adoptă servere ARM de înaltă densitate și consum redus de energie! 🚀#Proxmox #ProxmoxVE #ARM64 #Virtualization #Linux #OpenSource #KVM #ZFS #Ceph #SysAdmin #TechNews #FOSS

🖥️ Proxmox VE extinde oficial suportul dincolo de x86, adăugând arhitectura Arm64! Lansarea oficială a ediției Proxmox Virtual Environment (VE) 9.2 pentru Arm64 aduce suita completă de virtualizare pe servere bazate pe arhitectură ARM. Proiectul împărtășește același depozit de cod, funcționalități și ciclu de viață cu versiunea x86-64. ✨ Principalele noutăți și specificații tehnice:🚀 Paritate completă de funcționalități:• Oferă suport nativ pentru KVM, containere LXC, sistemele de fișiere/stocare ZFS și Ceph, clustering HA, backup și interfața web clasică de administrare. 🖥️ Hardware certificat & Suport Enterprise:• Optimizat pentru servere AI & Cloud: Certificare oficială din prima zi pentru platformele de servere NVIDIA Grace Hopper™ și NVIDIA Vera™.• Cerințe hardware strict server: Necesită boot prin UEFI și descrierea componentelor prin ACPI (compatibil cu ARMv9-A și ARMv8-A în regim best-effort). ⚠️ Fără plăci de dezvoltare (SBC):• Dispozitivele bazate exclusiv pe Device Tree (cum ar fi Raspberry Pi, Orange Pi etc.) NU sunt suportate oficial.⚙️ Stack tehnologic la bază:• Construiți pe Debian 13.5 "Trixie", kernel Linux 7.0, QEMU 11.0, LXC 7.0 și ZFS 2.4. 🔄 Diferențe specifice arhitecturii:• Mașinile virtuale pe Arm64 folosesc exclusiv AAVMF (UEFI) (SeaBIOS nu este disponibil).• Migrarea live a VM-urilor se poate face doar între noduri din aceeași arhitectură (x86 la x86, ARM la ARM). Un pas strategic masiv pentru datacenter-e și mediile enterprise care adoptă servere ARM de înaltă densitate și consum redus de energie! 🚀#Proxmox #ProxmoxVE #ARM64 #Virtualization #Linux #OpenSource #KVM #ZFS #Ceph #SysAdmin #TechNews #FOSS

Does Ceph just automatically ensure data replication across nodes for HA in event of a node failure or do I need to do something special?
I only have 2 nodes to test with and if I pull the cord on one, the other fails to bring the storage online, but I think that's just because Ceph can't function on only two nodes. In production we'd have a cluster of 4. Would it just seamlessly migrate any downed VMs on a failed cluster in that scenario?

#proxmox #ceph

Does Ceph just automatically ensure data replication across nodes for HA in event of a node failure or do I need to do something special?
I only have 2 nodes to test with and if I pull the cord on one, the other fails to bring the storage online, but I think that's just because Ceph can't function on only two nodes. In production we'd have a cluster of 4. Would it just seamlessly migrate any downed VMs on a failed cluster in that scenario?

#proxmox #ceph

#Ceph 20.2.3 has been released.

https://ceph.io/en/news/blog/2026/v20-2-3-tentacle-released/

Finally with PR #68751 to fix hybrid OSD deployments.

#Ceph 20.2.3 has been released.

https://ceph.io/en/news/blog/2026/v20-2-3-tentacle-released/

Finally with PR #68751 to fix hybrid OSD deployments.

Time for some Sunday Morning Homelab maintenance. First up: Ceph Upgrade to 2.20.2.

#HomeLab #Ceph

Time for some Sunday Morning Homelab maintenance. First up: Ceph Upgrade to 2.20.2.

#HomeLab #Ceph

Every time I look into Proxmox I find it I need another N+1 nodes.
First it was 3 nodes to maintain quorum
Then it was 5 to maintain resilience and keep an odd number.
Now I'm told in order to use Hyper-Convergence with Ceph I should really have 6 or more nodes and 100GB networking between them ​:blobcatnotlikethis2:​

#proxmox #ceph

Every time I look into Proxmox I find it I need another N+1 nodes.
First it was 3 nodes to maintain quorum
Then it was 5 to maintain resilience and keep an odd number.
Now I'm told in order to use Hyper-Convergence with Ceph I should really have 6 or more nodes and 100GB networking between them ​:blobcatnotlikethis2:​

#proxmox #ceph

Friendly reminder that https://tracker.ceph.com/issues/70390 was fixed in Ceph Squid 19.2.4 on 01. June 2026.

Full redeployment of affected OSDs is required (rm/zap) to fix the on-disk corruption. All OSDs created with Squid < 19.2.4 are affected. No manual fsck/repair possible.

#ceph #linux

An ein T im df -h habe ich mich ja schon gewöhnt, aber ein P ist noch ungewohnt für mich:

1.6P 447T 1.2P 29% /srv/IT

#ceph #sysadmin

An ein T im df -h habe ich mich ja schon gewöhnt, aber ein P ist noch ungewohnt für mich:

1.6P 447T 1.2P 29% /srv/IT

#ceph #sysadmin

Wir konvertieren gerade einen #Ceph-Cluster von croit zum Open-Source-Ceph mit Orchestrator im laufenden Betrieb.
So wie es aussieht, funktioniert der Plan. 😁

Wir konvertieren gerade einen #Ceph-Cluster von croit zum Open-Source-Ceph mit Orchestrator im laufenden Betrieb.
So wie es aussieht, funktioniert der Plan. 😁

Tá ficando pronto. E vai dar o que falar!

https://nuvem.online/blog/kubernetes-on-premises-infra-estatal-sem-lock-in

#soberaniaDigital #ibict #kubernetes #ceph #devops

A propósito: o avatar parece muito comigo, mas não sou eu. kkkk

Stop $30k GPUs from idling! ⚡ Eliminate GPU Starvation in AI training pipelines.

WekaFS (DPDK) vs Ceph on Bare Metal:

• WekaFS: Extreme DPDK speed, but high software licensing fees.
• Ceph: Zero lock-in, petabyte scale on 100G networks.
• IOMMU Fix: Use iommu=pt in GRUB to fix NVMe CPU spinlocks safely.
• K8s Mandate: Use CephFS (RWX), not RBD, for shared LLM weights.

https://www.servermo.com/howto/wekafs-vs-ceph-bare-metal/

#WekaFS #Ceph #AI #DevOps #Storage #ServerMO

Stop $30k GPUs from idling! ⚡ Eliminate GPU Starvation in AI training pipelines.

WekaFS (DPDK) vs Ceph on Bare Metal:

• WekaFS: Extreme DPDK speed, but high software licensing fees.
• Ceph: Zero lock-in, petabyte scale on 100G networks.
• IOMMU Fix: Use iommu=pt in GRUB to fix NVMe CPU spinlocks safely.
• K8s Mandate: Use CephFS (RWX), not RBD, for shared LLM weights.

https://www.servermo.com/howto/wekafs-vs-ceph-bare-metal/

#WekaFS #Ceph #AI #DevOps #Storage #ServerMO

Tausche #Netapp Cap meines Vorgängers gegen #Ceph oder #TrueNAS Cap 😁

Tausche #Netapp Cap meines Vorgängers gegen #Ceph oder #TrueNAS Cap 😁

Keine Marketing-Folien. Nur Proxmox. 🖥️

Der erste ProxDay – ein Tag voller Praxis rund um Proxmox & Virtualisierung.
Echte Talks, echtes Networking, von Profis für Profis.

📅 15. Oktober 2026
📍 Hotel Palace St. George, Mönchengladbach
🎟️ 35 € inkl. Catering & Getränke n vergeben

Jetzt Karte sichern 👉 https://www.proxday.de/

#Proxmox #Virtualisierung #ProxDay #Ceph #Kubernetes #Homelab #Linux #OpenSource

Keine Marketing-Folien. Nur Proxmox. 🖥️

Der erste ProxDay – ein Tag voller Praxis rund um Proxmox & Virtualisierung.
Echte Talks, echtes Networking, von Profis für Profis.

📅 15. Oktober 2026
📍 Hotel Palace St. George, Mönchengladbach
🎟️ 35 € inkl. Catering & Getränke n vergeben

Jetzt Karte sichern 👉 https://www.proxday.de/

#Proxmox #Virtualisierung #ProxDay #Ceph #Kubernetes #Homelab #Linux #OpenSource

Non-disruptive data migration between pools comes to #Ceph

There was a talk at the Ceph Day London from Tom Sollers and Bill Scales

This looks very promising and may land in Ceph 22.

Sildes on https://ceph.io/en/community/events/2026/ceph-days-london/
Video https://www.youtube.com/watch?v=pjTLCHxP6YA:

Non-disruptive data migration between pools comes to #Ceph

There was a talk at the Ceph Day London from Tom Sollers and Bill Scales

This looks very promising and may land in Ceph 22.

Sildes on https://ceph.io/en/community/events/2026/ceph-days-london/
Video https://www.youtube.com/watch?v=pjTLCHxP6YA:

Cloning a terabyte production environment in under a minute sounds like engineering magic, but it is just Upsun leveraging Ceph copy-on-write. ⚡
However, every great shortcut comes with a catch, like a 16-level clone-depth limit hiding right behind it.

Dive into how a one-minute clone actually works and how to walk the chain without breaking your architecture. 🛠️

👉 https://developer.upsun.com/posts/how-it-works/cloning-a-terabyte-in-a-minute-and-the-limit-hiding-behind-it

#CloudNative #Ceph #DevOps #Storage

Cloning a terabyte production environment in under a minute sounds like engineering magic, but it is just Upsun leveraging Ceph copy-on-write. ⚡
However, every great shortcut comes with a catch, like a 16-level clone-depth limit hiding right behind it.

Dive into how a one-minute clone actually works and how to walk the chain without breaking your architecture. 🛠️

👉 https://developer.upsun.com/posts/how-it-works/cloning-a-terabyte-in-a-minute-and-the-limit-hiding-behind-it

#CloudNative #Ceph #DevOps #Storage

Wir messen hier grad 1,4 Millionen read IOPS mit 4k auf einem Ceph-Cluster. 16 fio Prozesse mit iodepth=256. 16 Ceph-Knoten mit je 6 NVMe und 2x 100G Netzwerk.
4k schreibend sind es immer noch ca 100.000 IOPS. 4M schreibend erhalten wir ca 60GB/s.
Alles auf einem 10+3 EC Pool mit den FastEC-Optimierungen aus #Ceph 20. Nice.

Wir messen hier grad 1,4 Millionen read IOPS mit 4k auf einem Ceph-Cluster. 16 fio Prozesse mit iodepth=256. 16 Ceph-Knoten mit je 6 NVMe und 2x 100G Netzwerk.
4k schreibend sind es immer noch ca 100.000 IOPS. 4M schreibend erhalten wir ca 60GB/s.
Alles auf einem 10+3 EC Pool mit den FastEC-Optimierungen aus #Ceph 20. Nice.

Weird Ceph puzzle (Tentacle 20.2.2, kernel CephFS client, EC 4+2 pool, BlueStore/HDD):
fio sequential write, O_DIRECT, iodepth=32, numjobs=16:

bs=4M -> ~3.3 GB/s
bs=32M -> ~1.8 GB/s (nearly half!)
Same everything else. Reads unaffected either way.

Ruled out so far:

CephFS object_size=4M, so a 32M write spans 8 objects. Tried stripe_count=8 for real client-side striping across objects -> no change.
Tried io_uring vs libaio. io_submit() latency dropped ~360x (283ms -> 0.8ms avg), confirming libaio was blocking on submission -> but aggregate bandwidth stayed identical (1890 vs 1889 MB/s). So AIO engine isn't it either.
wsize/rsize already at the kernel client's real max (64MB - yes, exceeding it hard-fails the mount with EINVAL).

Reproducible across 5 runs, 2 different EC stripe_unit pools, same magnitude both times.
Anyone seen this / have a theory?

#ceph #linux #storage #sysadmin

Weird Ceph puzzle (Tentacle 20.2.2, kernel CephFS client, EC 4+2 pool, BlueStore/HDD):
fio sequential write, O_DIRECT, iodepth=32, numjobs=16:

bs=4M -> ~3.3 GB/s
bs=32M -> ~1.8 GB/s (nearly half!)
Same everything else. Reads unaffected either way.

Ruled out so far:

CephFS object_size=4M, so a 32M write spans 8 objects. Tried stripe_count=8 for real client-side striping across objects -> no change.
Tried io_uring vs libaio. io_submit() latency dropped ~360x (283ms -> 0.8ms avg), confirming libaio was blocking on submission -> but aggregate bandwidth stayed identical (1890 vs 1889 MB/s). So AIO engine isn't it either.
wsize/rsize already at the kernel client's real max (64MB - yes, exceeding it hard-fails the mount with EINVAL).

Reproducible across 5 runs, 2 different EC stripe_unit pools, same magnitude both times.
Anyone seen this / have a theory?

#ceph #linux #storage #sysadmin

#ceph orch upgrade start --image quay.io/ceph/ceph:v20.2.2
macht einfach mal das, was man erwartet einfach so. 🥳

If you're upgrading #Ceph on #Proxmox VE v8 AKA Debian/bookworm, be aware of the ceph init script situation in the ceph-base package. Just tracked this down + reported it towards upstream:

https://bugzilla.proxmox.com/show_bug.cgi?id=7777

(Nice bug number! 🤓)

If you're upgrading #Ceph on #Proxmox VE v8 AKA Debian/bookworm, be aware of the ceph init script situation in the ceph-base package. Just tracked this down + reported it towards upstream:

https://bugzilla.proxmox.com/show_bug.cgi?id=7777

(Nice bug number! 🤓)

I present to you, my Ceph cluster after I issued a parallel "crictl rmi --prune" on all of my k8s workers. I didn't even know my cluster could reach 600 MB/s client throughput. 😅

#HomeLab #Ceph

I present to you, my Ceph cluster after I issued a parallel "crictl rmi --prune" on all of my k8s workers. I didn't even know my cluster could reach 600 MB/s client throughput. 😅

#HomeLab #Ceph

Accelerate slow HDD-backed OSDs in Proxmox by creating a Ceph storage pool with SSD cache tiering. This snippet configures a hot tier (SSD) with target size and hit set count to promote frequently accessed objects automatically.

#proxmox #ceph #cache-tiering

https://www.valtersit.com/vault/ceph-storage-pool-with-ssd-cache-tiering-88d504/

#ceph cluster upgrade from 16 to 17 to 18 to 19 to 20 successful :catjam:

#ceph cluster upgrade from 16 to 17 to 18 to 19 to 20 successful :catjam:

I looked at #Ceph code today, #AMA.

( https://github.com/ceph/ceph )

I looked at #Ceph code today, #AMA.

( https://github.com/ceph/ceph )

These are the guilty culprits… one just replaced now, and I'll replace the other soonish.

Right now #Ceph is getting re-acquainted with the data on the node I just brought back up today, and I'll wait for that to finish before I shut down the other node which is currently running on a temporary drive. I don't want to rug-pull the recovery.

The replacements are these things:
https://www.umart.com.au/product/silicon-power-a55-128gb-m-2-2280-sata-iii-ssd-3d-nand-with-slc-cache-speed-560mb-s-for-desktop-laptop-sp128gbss3a55m28-61639

That's all I can get in a hurry… don't trust them as far as I can throw them, but for now they get the machines booting. I might put an order into RS and get some Kingston M.2 SATA SSDs, the mSATA versions have been fine so far in the machines I've put them in and they're not ridiculously expensive.

These are the guilty culprits… one just replaced now, and I'll replace the other soonish.

Right now #Ceph is getting re-acquainted with the data on the node I just brought back up today, and I'll wait for that to finish before I shut down the other node which is currently running on a temporary drive. I don't want to rug-pull the recovery.

The replacements are these things:
https://www.umart.com.au/product/silicon-power-a55-128gb-m-2-2280-sata-iii-ssd-3d-nand-with-slc-cache-speed-560mb-s-for-desktop-laptop-sp128gbss3a55m28-61639

That's all I can get in a hurry… don't trust them as far as I can throw them, but for now they get the machines booting. I might put an order into RS and get some Kingston M.2 SATA SSDs, the mSATA versions have been fine so far in the machines I've put them in and they're not ridiculously expensive.

Well, you picked a fine time to leave me boot disk…
Along with your comrade, blowing my OS away…
My pair of downed servers, missed 6TB of replicas…
now my Ceph cluster won't heal…
…you picked a fine time to leave me, boot disk!

#HomeServer #Ceph

Well, you picked a fine time to leave me boot disk…
Along with your comrade, blowing my OS away…
My pair of downed servers, missed 6TB of replicas…
now my Ceph cluster won't heal…
…you picked a fine time to leave me, boot disk!

#HomeServer #Ceph

[ Blog ] Proxmox upgrade #Ceph #Reef to #Squid

If you are running Ceph Reef in your #Proxmox infrastructure and plan to upgrade to Proxmox 9, you must first upgrade Ceph Reef to Squid to meet the prerequisites.

As a best practice, before proceeding with the upgrade make sure to have a working backup of your VMs and Containers.

 
Prerequisites
To upgrade Ceph Reef to http://rviv.ly/tmWeXQ #aggiornamento

[ Blog ] Proxmox upgrade #Ceph #Reef to #Squid

If you are running Ceph Reef in your #Proxmox infrastructure and plan to upgrade to Proxmox 9, you must first upgrade Ceph Reef to Squid to meet the prerequisites.

As a best practice, before proceeding with the upgrade make sure to have a working backup of your VMs and Containers.

 
Prerequisites
To upgrade Ceph Reef to http://rviv.ly/tmWeXQ #aggiornamento

Alle, die besonders großen Speicherbedarf haben, kommen im Open-Source-Bereich an #Ceph als #StorageLösung eigentlich nicht vorbei. Ceph eignet sich für verschiedenste Einsatzszenarien – vom Daten-Backup bis zum Ersatz für Amazon S3 - und bringt dazu bereits gute Lösungen mit.

Aber das passende Ceph-Cluster für die individuellen Zwecke aufzubauen ist komplex. Hier helfen unsere Profis weiter:

Ceph-Kurs an der #HeinleinAkademie: https://www.heinlein-support.de/schulung/ceph-grundlagen

Ceph-Beratung: https://www.heinlein-support.de/consulting/ceph

Alle, die besonders großen Speicherbedarf haben, kommen im Open-Source-Bereich an #Ceph als #StorageLösung eigentlich nicht vorbei. Ceph eignet sich für verschiedenste Einsatzszenarien – vom Daten-Backup bis zum Ersatz für Amazon S3 - und bringt dazu bereits gute Lösungen mit.

Aber das passende Ceph-Cluster für die individuellen Zwecke aufzubauen ist komplex. Hier helfen unsere Profis weiter:

Ceph-Kurs an der #HeinleinAkademie: https://www.heinlein-support.de/schulung/ceph-grundlagen

Ceph-Beratung: https://www.heinlein-support.de/consulting/ceph

Легаси-ОС как тормоз виртуализации: что меняет современный стек РЕД ОС в VK Cloud

Представьте гиперноду облака. Гипернода — это физический сервер с запущенным гипервизором, на котором работают виртуальные машины клиентов. Под дисками этих машин лежит программно определяемое хранилище Ceph: распределенная система, где данные размазаны по многим серверам с копиями, без отдельного дискового массива. Меняем на ноде одну переменную — операционную систему. Виртуальные машины не пересобираем, кластер хранения не трогаем, диски и сеть те же. Ни одной новой железки, ни строчки нового кода в приложении. После переключения дисковая подсистема ВМ ведет себя ощутимо иначе. VK Cloud активно использует РЕД ОС от РЕД СОФТ — в том числе в VK Secure Cloud, аттестованном контуре для значимых объектов критической информационной инфраструктуры (ЗОКИИ). На ее примере покажу, как поднять производительность гипервизора, просто обновив легаси и не трогая железо. Вместе с дистрибутивом на ноду приезжает свежий стек целиком: ядро, эмулятор, клиент хранилища, системные библиотеки. Каждый слой подтягивает свой кусок. А для тех, кто застрял на CentOS, ушедшем в EOL, у истории есть вторая часть: обновление закрывает технический разрыв и регуляторику одним движением. Ниже разберу механику по слоям с командами, которые можно выполнить на своей системе.

https://habr.com/ru/companies/vktech/articles/1046534/

#vk_cloud #vk_secure_cloud #red_os #centos_eol #ceph #kvm #qemu #io_uring #импортозамещение #vk_tech

I decided to do a series of posts about my #homelab architecture.

I'm not quite done with the series yet! I still need to detail the roadmap I currently have planned, but it's complete enough at the moment to post.

Enjoy!

https://crabby.vellas.ca/posts/homelab-architecture-overview/

#proxmox #ceph #selfhosted

I decided to do a series of posts about my #homelab architecture.

I'm not quite done with the series yet! I still need to detail the roadmap I currently have planned, but it's complete enough at the moment to post.

Enjoy!

https://crabby.vellas.ca/posts/homelab-architecture-overview/

#proxmox #ceph #selfhosted

PSA: The Ceph Rook update to 1.20 is a bit more involved due to the introduction of the ceph-csi-drivers Helm chart. Have a look at the upgrade docs: https://rook.github.io/docs/rook/v1.20/Upgrade/rook-upgrade/

Also have a look at this GitHub issue for a few things which might be helpful but haven't made their way into the docs yet: https://github.com/rook/rook/issues/17696

#HomeLab #Ceph #CephRook

Sommerpause mit Linux-Wissen: Die Heinlein Akademie gönnt sich ab Mitte Juli 2026 eine kurze Auszeit. Die gute Nachricht: Du musst trotzdem nicht auf maßgeschneiderte Linux-Schulungen verzichten.

Neben dem regulären Kursangebot bieten wir Teams #InhouseSchulungen, die wir gemeinsam individualisieren – egal ob Dauer, inhaltlicher Schwerpunkt, Ort oder Termin.

Dieses Jahr haben unsere Profis z.B. Teams zu #Kubernetes, #Proxmox, oder #Ceph geschult.

Wir beraten gerne:
👉 https://www.heinlein-support.de/akademie/inhouse-schulungen

[ Blog ] Proxmox upgrade #Ceph #Reef to #Squid

If you are running Ceph Reef in your #Proxmox infrastructure and plan to upgrade to Proxmox 9, you must first upgrade Ceph Reef to Squid to meet the prerequisites.

As a best practice, before proceeding with the upgrade make sure to have a working backup of your VMs and Containers.

 
Prerequisites
To upgrade Ceph Reef to http://rviv.ly/tmWeXQ #aggiornamento

On a related note:

Does anyone know good beginner-level documentation on the getting the actual configuration going? So once you have Ceph or OpenStack up and running, how do you organize the thing?

For Ceph, this means pools and whatnot.

For OpenStack, this means domains and projects and networks. Which ones should be created by the admin, which ones should be up to the actual users (in their tenant aka project)? Best practices for roles, permissions, things like that.

The docu for both Ceph and OpenStack is excessive, but very detailled or very high-level. I would like to see a hands-on guide that does not just do some examples, but has recommendations or pros & cons or the like.

#OpenStack #Ceph #homelab #kollaAnsible #kolla #HellYeah #WhyNot @homelab

remember folks, #ceph is not a microservice