Hello!
You do not need any optimizations here since the time between checkpoints
is configurable.
More checkpoints - faster recovery but more data needs to be written to
disk.
Fewer checkpoints - recovery will take more time but overall performance is
somewhat higher.
Regards,
--
Ilya
Ilya,
Does 2.9 have specific optimizations for checkpoints?
Thanks,
Raymond.
On Fri, Oct 9, 2020 at 1:23 AM Ilya Kasnacheev
wrote:
> Hello!
>
> I think the real saver is in decreasing amount of time between checkpoints.
>
> Regards,
> --
> Ilya Kasnacheev
>
>
> вт, 6 окт. 2020 г. в 02:10,
Hello!
I think the real saver is in decreasing amount of time between checkpoints.
Regards,
--
Ilya Kasnacheev
вт, 6 окт. 2020 г. в 02:10, Raymond Wilson :
> Thanks for the thoughts Ilya and Vladimir.
>
> We'll do a comparison with 2.9 when it releases to see if that makes any
> difference.
Thanks for the thoughts Ilya and Vladimir.
We'll do a comparison with 2.9 when it releases to see if that makes any
difference.
One of the advantages with persistent storage is that it is effectively
'instant start'. Our WAL size is around 5Gb, perhaps this should be
decreased to reduce system
It's possible that it happens because of
https://issues.apache.org/jira/browse/IGNITE-13068.
We need to scan the entire SQL primary index during startup in case you have
at least on query entity configured.
As far as I can see it's going to be a part of the Ignite 2.9 release.
--
Sent from:
Hello!
I guess it is the time to iterate WAL after the last checkpoint to make
sure all operations prior to node stop are accounted for.
Every DB will do that and the time of 76 seconds seems reasonable.
Regards,
--
Ilya Kasnacheev
вт, 29 сент. 2020 г. в 02:00, Raymond Wilson :
> There is a
There is a summary log line that also references the time taken by the
'restore binary memory' stage. It also references another stage called
'restore logical state'.
eg:
2020-09-27 20:50:31,423 [1] INF [] Node started : [stage="Configure system
pool" (44 ms),stage="Start managers" (2422
Just a general query on Ignite server node startup sequence.
There are a pair of log lines emitted during node startup that look like
this:
2020-09-27 20:48:43,191 [1] INF [] Starting binary memory restore for:
[948695530, -1026154091, -2049124072, -552847479, -2024686410, -1820754607,