Hey guys!
We have a big grid, where we decided to change EC. Since its such a big grid and a lot of data we do the change in batches. Creating a new rule to change data that was ingested in a specific time range, to not put to much strain on the grid.
However do anyone know if its possible to track the progress somehow? So we can get an estimate when the job is completed.
#Can you track your progress after changing EC?
lime coral
tight drum
Hi @lime coral there is some graphing of the ILM activities, have a look at https://docs.netapp.com/us-en/storagegrid-116/monitor/monitoring-information-lifecycle-management.html
I don't think you get a specific ETA though.
lime coral
Hi <@214434201141903360> there is some graphing of the ILM activities, have a lo...
Yeah, I'm aware about that graph. I Was hoping that someone had some trick up their sleeves to see a rough estimate 🙂
Thanks anyway thou!
But maybe you can help me with another quick question. When looking at graphs on the load balancer e.g. Nodes > site name > Load balancer you see operations for DELETE, GET, HEAD, POST, PUT.
May I ask what POST is? It started to spike here and there after we implented the new rule. I have no clue what exactly POST is thou.
tight drum
Are you using a Cloud Storage pool within your grid?
lime coral
Are you using a Cloud Storage pool within your grid?
No, we do not use Cloud Storage pool
final parcel
here is some info that might provide additional insight
extract from links
If an object is stored in a Cloud Storage Pool and one or more copies of the object also exist on the grid, there is no need to restore the object by issuing a POST Object restore request. Instead, the local copy can be retrieved directly, using a GET Object request.
is the new rule using the same EC profile or a new profile
either ways, new rule is probably seeing multiple copies of the data and is invoking a POST operation while redistribute data per the new rule
lime coral
is the new rule using the same EC profile or a new profile
We move from 6+1 to 8+2
lime coral
If an object is stored in a Cloud Storage Pool and one or more copies of the obj...
Not sure if I understand...
Since we don't use Cloud Storage Pool at all. Why does it show high latency on POST? Shouldn't ONLY a GET be requested?
final parcel
I assume there isn't any other jobs running as in there aren't any backup, restores or anything else going on related to the GRID
if so, can you check your Network activity, does your network graph show busy traffic from the time new ILM was applied
final parcel
going from 6+1 to 8+2 requires data to be reconstructed to whole first before applying 8+2. Reconstruction of data requires a partial restore process to be initiated to a point that data is whole before breaking it down to 8+2 which is why you are seeing POST as opposed to GET
Network activity going up soon as new rule was initiated is the network activity between the storage related to reconstruct of data and then distribute it per the new rule