IDE died while editing a node

[sylwiabr]

Discord username

No response

What type of issue is this?

Intermittent – Occurring irregularly

Is this issue blocking you from using Enso?

• Yes, I can't use Enso because of this issue.

Is this a regression?

• Yes, previous version of Enso did not have this issue.

What issue are you facing?

It happened few times during last two days: IDE died without any message to the user while working normally (editing the node). The error in the console was: Peer has replied with an error: File not found . Full session had just around 4 min so attaching the link to the video: https://drive.google.com/file/d/1ajtJTaEkFsGMe0bg8YnO5SAt9aGGtf7x/view?usp=share_link

Expected behaviour

IDE should never die.

How we can reproduce it?

No response

Screenshots or screencasts

No response

Enso Version

develop 12.05

Browser or standalone distribution

Standalone distribution

Browser Version or standalone distribution

standalone

Operating System

Windows

Operating System Version

No response

Hardware you are using

No response

[sylwiabr]

@farmaazon @4e6 have a look here. IDE should never die - it was a short session, nothing controversial. During last days we encountered exactly this error 2-3 times and we had to restart IDE multiple time because it was unusable. Also we lost our work few times due to this and other problems (like problems with synchronization between IDE and engine).

[jdunkerley]

I think making the engine sluggish by running in trace mode will help reproduce.

Project: https://drive.google.com/file/d/11jOQvmkEhplUfPDlvGPnESkiV5EXnmoO/view?usp=share_link

[JaroslavTulach]

The video shows following log at 3:42:

At 22:06 the HeartbeatSession] stopped. At 22:15 the HeartbeatSession] started - what does that mean? Is the IDE trying to reconnect? Btw. at 22:21 the heartbeat stops again.

Is there any other explanation of the heartbeat stopping/starting than an attempt to reconnect? The IDE logs (shown at 4:02 of the video) do not contain any message about reconnecting, however.

[farmaazon]

The Heartbeat is an internal Engine thing, IDE do not use it.

[hubertp]

@JaroslavTulach Heartbeat session is a shortlived action that starts and dies regularly. sending/receiving the regular ping/pong. This is as expected.

[JaroslavTulach]

sending/receiving the regular ping/pong. This is as expected

That means the connection between the IDE and engine is OK, right? Alas, if it is OK, then I have no other idea what can be wrong...

[sylwiabr]

@JaroslavTulach the issues we are observing recently like this one, #6727 or #6634 are not caused by IDE/LS loosing the connection. Probably this is why IDE is not reporting issues. They all have common part: evaluation of newly added nodes is not happening but the connection is there.

[JaroslavTulach]

We need to find out what is happening in the Enso engine/LS process - let's download VisualVM and then:

• open the org.enso.runner.Main process
• take a thread dump (Threads tab/Thread Dump button / copy and paste)
• sample what the application does (Sampler tab / CPU button / Snapshot / save into a file)
• take a heap dump (right click on the process node)

With such information we should have a better insight of what is happening in the engine/LS process:

• thread dump would show us if the engine is blocked somewhere
• profiling sample allows us to record few seconds and deduce what the engine is doing at that time
• heap dump allows us to inspect internal objects like actors, etc.

[JaroslavTulach]

Alas, we have just done both (thread dump and profiling) with James and the LS/engine process does nothing! The next step is to find out whether there are some messages sent to the process at all.

[hubertp]

Sounds vaguely similar to #6763

[JaroslavTulach]

I guess we should invest into an infrastructure that would allow us to capture/record the state of the engine easily:

• Simplify taking a snapshot/sampling with an IDE shortcut #6775

[jdunkerley]

Closing as we believe is related to #6763.