Update known issues

docs/release/knownissues.md

@@ -15,3 +15,25 @@ One option might be to switch the streaming mode to "Jpeg", however, we'd recomm
 <p align="center">
   <img width="70%" src="assets/videostream.png">
 </p>
+
+### 2. Job submission into the cloud
+
+There is currently a bug that prevents the user to have 10 concurrent iSolve jobs running. If the jobs are submitted within a short time of each other, they will eventually end up on one ec2-instance and run sequentially. This typically happens with MultiPort Simualtions.
+
+### 3. Integer buffer overflow with MPI iSolve Jobs
+
+Large LF simulations that use `PETSc` can fail with integer overflow. The workaround is to use more than 1 or 2 MPI ranks in the Solver Settings.
+
+### 4. Cuda simulation memory error
+
+The EmFdtd Cuda solver suffers from a memory allocation buffer. This typically happens when simulations run on more than on NVIDIA A10 cards (`MEDIUM`). Workaround is to use the other option with hardware acceleration.
+
+### 5. Disk space limits
+
+The current limits for disk space are for the sim4life GUI are:
+
+- `MEDIUM`: 200 GB
+- `LARGE`: 800 GB
+- `XLARGE`: 800 GB
+
+Due to the algorithm used to orchestrate the machines, it may happen that you end up on a `LARGE` machine (with constraint CPU/RAM usage) although you chose a `MEDIUM`. Since we do not yet limit the disk space you might actually create a 200 GB workspace without noticing. However, when you start the project the next time on a `MEDIUM` machine, you will run out of disk space. (You can see this in the log of a failed to start service). Workaround in that case is to use a `LARGE` instance for your project.