Linger | Simple picture sharing system

I am writing a small TCP sockets "library", and I ran into much trouble.

When something happens to the socket that causes it to be instantly closed and freed (regardless of background lingering, talking about user space here), which is only done within a locked mutex since the application I'm writing is multi-threaded, I need a way to tell all the other (potentially) waiting threads (on the same mutex) that want to do something with the socket: "I'm terribly sorry, but the socket has been destroyed and you cannot access it" so that they don't cause any segmentation fault or such.

The idea I had was: the mutex was part of the socket (socket = a structure that contains multiple things, including a mutex and a file descriptor) so I couldn't quite free the socket if other threads were waiting for it (undefined behavior), so the possible solution is to allocate the mutex, free the socket but not the mutex, set some flag (in the allocated memory) saying that the socket has been closed, and a counter so that the last thread waking up and getting notified that it cannot use the socket unlocks and destroys the mutex and frees the memory allocated. The mutex can still be accessed without segfault if we just store its pointer before acquiring the mutex (and the pointer won't ever change).

This solution has a fundamental problem though - what if between unlocking the mutex and freeing it by the last holder, the thread gets preempted and another one locks the mutex again (since you NEED to check socket-related stuff after you acquire the lock, so no way of knowing it got destroyed, unless you maybe use an atomic variable but then again, checking the atomic variable and locking the mutex as a whole are not an atomic operation). Or if you try to access the mutex on the already-freed socket. Undefined behavior emerges.

Any ideas how to solve this problem? I.e. how to destroy a socket so that other threads know about it to quit safely, and there are no race conditions? By quitting I mean aborting the socket function they were in, not cancelling or stopping the threads themselves.

I am developing a chrome extension for the new tab page and I am trying to find the right DB for the project. The only question that is keeping me from using Firebase Firestore is to know how the DB handles reads.

Basically, every time the user opens a new tab page I will need to fetch around 3000 (very small) documents (hopefully from cache). My issue is that since opening a new tab page is done so frequently I will be charged an absurd amount of reads because firestore is always reading 3K documents.

My question is, is Firestore smart enough to tell that in the DB data has not changed and the client should only read from the cache?

I read all about offline persistence but this question is still lingering!

Thank you for any help!

Okay, the title is quite mouthful. But it's actually describing the situation.

I deployed a service on GKE in namespace argo-events. Something was wrong with it so I tore it down:

In a mobile app that I am currently developing I keep track, serverside, of how many visitors enter/leave the area of visibility of one or more BLE beacons. The app communicates the unique (string) ID of the beacon along with the day (as a difference from 01/04/2021) and 5 minute time slot at which the event happened. All of this ends up - via an intermediate Redis clearing house - in the following PGSQL table

I keep getting 403 forbidden access error when trying to access any project within htdocs folder. I even downloaded a vanilla CodeIgniter 3 project and got the same result. I'm using apache 2.4.

Above are my conf files.

httpd.conf:

I have an app I am developing and the stakeholder using it said that the app becomes slow and unusable/unresponsive after consistent usage all day. Killing it and starting over causes it to run fine.

I don't seem to have this trouble on my device, but I started looking at the memory usage in both simulator/phone in debugger, and observed my memory would steadily increase if I took the basic action of going between screen to screen. These are pretty involved screens, but if I just go forward to the 'add new item' screen, then back to the product listing screen, the memory jumps up 30mb. If I keep doing this same action, over and over and over, I can get it to 1.1gb of memory

I then took it a step further, hooked up my phone, and ran profiler (specifically memory leaks). I found one leak involving my usage of ads, so I just commented out all the code for a test and while the leaks are gone, the memory continues to go up steadily.

I then ran the allocations tool, and after a few min of going back and forth in the same manner, here is the output:

As you can see, it's 1.53GB and if I kept doing the same action I can get it to 2GB+. Oddly enough, my phone never seems to mind, and the screens are just slightly laggy at times otherwise not too bad. Certainly usable.

Before I start ripping out the floor boards, I wanted to confirm this is a likely sign of a problem. Any suggestions on where I can start looking? If persistent memory is the issue, what would be some typical gotchas or pitfalls? What is "anonymous vm?"

Thank you so much if you're reading this far, and appreciate any guidance!

UPDATE/EDIT

After some guidance here, I noticed, oddly enough, that on the "add product" page it causes the memory to jump ~10MB each time I visit it. After commenting out code, I narrowed it down to this section (and even the line of code) causing the jump. Removing this code causes it to remain stable and not increase.

After compiling I receive this error message:

Failed to compile src\App.js Line 4:1: 'state' is not defined no-undef

Code App.js:

Versions:

• Spring boot: 2.4.1
• Kafka-clients: 2.6.0

Due some system limitations Kafka Producer is configured to send each message separately. Specifically, batches are disabled by setting linger.ms = 0 and max.batch.size = 0. Therefore, each message is new batch from Kafka's client perspective. I want to distribute messages equally between all available partitions, that's why I configured client RoundRobinPartitioner to achieve it.

However, testing on configuration when one node is down and I have even number of partitions, I figured out that messages are distributed between half of all available partitions. The reason for such behaviour is double invocation of partitioner.partition(..) in KafkaProducer.doSend(..)'. Since RoundRobinPartitioner.partition() under the hood increments counter and returns its remainder of the division by available partitions number, calling it twice within one record publish causes skipping of each second partition.

For instance, availablePartitions contains 6 partitions (1-6). Calling partition(..) twice through KafkaProducer.doSend(..) always skip 1,3,5 partitions.

I'm trying to tranfer data from Kafka topic to Postgres using JDBCSinkConnector. After all manipulations such as creating the topic, creating the stream, creating sink connector with configuration and produce data into topic throught python - connect logs returns the following result: