←back to thread

5 points raydenvm | 1 comments | | HN request time: 0.223s | source

I'm thinking about setting up a MiniNAS for our small development team. Mainly, for local storage of large artifacts and development builds.

I would love to hear about your real experiences with these compact NAS solutions. Any unexpected issues or limitations you discovered after you have started using one of them?

1. Bender ◴[] No.44472402[source]
I have not built a mini-nas for team use and personally would not from my experience with home grown solutions. I've worked with people that did things like this for cost cutting measures but it can end up costing millions or much more if work flows are interrupted. At a very minimum there would need to be n+1 live mini-nas with active replication to the stand-by nodes and there would need to be snapshots and backups. There would need to be really good monitoring of everything. These devices do not have the physical expansion capacity to do this correctly for large amounts of data in my opinion after snapshots. It may start off fine but storage requirements will only go up. Capacity planning with mini-nas constraints will quickly paint one into a corner. Mini computers also have heat efficiency issues that would be exacerbated by multiple people accessing them with heavy IO automation tools that you might not be using now but your team may in the future. There are ways to mitigate thermal issues but then the advantages of a mini-form factor would quickly go out the window. Another issue is power backup. Corporate systems deemed critical will be on UPS and generators. Most datacenters will not permit home grown solutions especially when they do not have redundant power inputs and assorted certifications that reduce the risk these devices will trigger fire suppression systems.

Now mini-NAS for backups of backups? That would make sense to me if you do not trust whomever is managing your administrative and backup servers. It might not even be full backups but one could at least back up the data critical to the business that your team is responsible for. A development team lead could automate more frequent backups to their mini-nas than the company has implemented for the wider audience use cases. That would not even need to be in the data-center unless your privacy, compliance and security teams say otherwise. Your home grown solution will still need really good physical security as random contract janitor can just walk out of the building with all of your intellectual property. The NAS just needs a very fast low latency connection to the data-center.

In my experience the revenue impacting data storage and flows should always be on the corporate maintained infrastructure unless one really wants to stand in front of the C-levels explaining why all the data was lost of a home grown implementation. Ideological and technical issues aside the optics will be awful. I've seen people walked out of companies for much less. Augmenting the corporate systems on the other hand can be a life saver especially if you know what data is critical to revenue flows and how many snapshots and full backups would keep your teams workflows uninterrupted. As an augmentation your systems could save the day and your team would look really sharp. As a side note, when your team does save the day by going above and beyond ensure that your management write up your team for awards. That can reduce management friction in the future and buy some leniency for mistakes that will inevitably occur.