You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Oct 30, 2018. It is now read-only.
Replace the values below using the output from storjshare --version.
* daemon: 2.1.0, core: 6.0.13, protocol: 1.1.0
Replace the values below using the output from node --version.
6.9.2
Expected Behavior
Startup speeds should be similar between the old CLI and the current Daemon approach.
Actual Behavior
I use to be able to start up 1160 nodes in slight under 20 minutes, simply by invoking a startup in backgroun followed by a 1 second sleep (in order to avoid trashing the OS). All nodes would then run as expected.
Using the daemon on the same system, each startup takes 8-9 seconds, requiring over 3 hours to spin up all 1160 nodes. Attempts to start in parallel, by running the command in background and spawning them 1 second apart like what was done with the CLI, appears to work, although nodes get started out of order and still apparently only one at a time, 8-9 seconds apart - by the time all where done, over 500 had disappeared
Steps to Reproduce
Please include the steps the reproduce the issue, numbered below. Include as
much detail as possible.
...
...
...
The text was updated successfully, but these errors were encountered:
Ah, so reason it is taking so long is because each command is establishing a new connection to the daemon, sending RPC, daemon receives and validates config, forks share process, responds to RPC, then closes connection. There is also a short artificial delay between the beggining of this and end to ensure the connection to the daemon is opened and closed cleanly.
The best way to start a series of nodes is to get them all running as you like and then running storjshare save to create a snapshot. This will let you later resurrect all the shares with one command storjshare load which eliminates the latency added by running each command separately. It will fork N shares with a degree of parallelism until all are running.
Package Versions
Replace the values below using the output from
storjshare --version
.Replace the values below using the output from
node --version
.Expected Behavior
Startup speeds should be similar between the old CLI and the current Daemon approach.
Actual Behavior
I use to be able to start up 1160 nodes in slight under 20 minutes, simply by invoking a startup in backgroun followed by a 1 second sleep (in order to avoid trashing the OS). All nodes would then run as expected.
Using the daemon on the same system, each startup takes 8-9 seconds, requiring over 3 hours to spin up all 1160 nodes. Attempts to start in parallel, by running the command in background and spawning them 1 second apart like what was done with the CLI, appears to work, although nodes get started out of order and still apparently only one at a time, 8-9 seconds apart - by the time all where done, over 500 had disappeared
Steps to Reproduce
Please include the steps the reproduce the issue, numbered below. Include as
much detail as possible.
The text was updated successfully, but these errors were encountered: