Hi @MWestfall
If you are offline then it will be counted against the node, however the 4-6 hours you expect is not long and completely permitted within the SNO terms and conditions. Bring the node back online as soon as you are able to not incur deletions from repair, for the data you hold.
if you get back online before the 4hour mark the network won’t even start to repair your data, and thus nothing will get trashed on your node… but i’ve at times had days of downtime… its really not to bad, ofc its best to limit it within the 4 hour before repair and never go below 60% online score, but that should take about 12 days.
Well I’ve got to splice some fiber, lol so /fingerscrossed
But it sounds like the ohh shit point is about 4 hours. After 4 hours, some of the data on my node may become invalidated and be repaired to other nodes. At which point when my node checks back in, it will then discard that data?
Is this done opportunistically? For example, if 6 hours is within acceptable “SNO” downtime, then that means no adverse actions should be taken. However, I completely understand that say a node is down. And an end user accesses their data, StorJ feels compelled to ‘remirror’ that data if your node is down even if others storing it are not, because they have no way of knowing if your node is down for 4 5 6 24, or forever hours.
And since that user just accessed their data it is their obligation to keep their users’ data ‘healthy’ thus triggering a repair/remirror of the recently accessed data.
Similarly, let’s say there’s a data block A that exists on say 4 nodes. And two nodes report offline on that same data block A. At that point, redundancy is seriously compromised even if the node has only been down for say 15 minutes. Would StorJ remirror/repair that data if multiple nodes storing data block A have been down for any period of time?
it’s nothing to worry about like i said i’ve been down for days at times, it happens no matter how hard we try.
its a couple of % of the total usually, if you even notice it
i would think it gets worse the longer one is down and it seems to me that if one is often down the network might already have enough that it won’t even attempt to repair.
but i duno the details, thats how it seems to me from what i’ve seen.
I would suggest you to leverage this “opportunity” to perform a file system check while you’re offline, so at least you do something useful which can be performed only when the node is shut down.
There is no special function for that, you just stop the node and run it back ASAP when your maintenance is finished. Your node can be offline up to 12 days, then it will be suspended (no ingress), then you will have a week to fix an issue, after the issue is fixed, your node will be on review for the next 30 days. To fully recover your node should be online for the next 30 days.
If your node would be offline more than 30 days, it will be disqualified.
Each downtime requires another 30 days to recover.
While your node is offline, its pieces could be recovered to other nodes, so when you bring your node online, the garbage collector will remove these pieces from your node.
In another words - more your node offline, more pieces it will lost.