I copied your comment to the issue thread. Maybe it is best to hold the conversation there as to leave a better trace.
On 13 Oct 2021, at 13:29, Henrik Sperre Johansen <henrik.s.johansen@veloxit.no> wrote:
Seens it used to be in the «old» PharoVM before the reconcilliation: https://forum.world.st/external-semaphores-again-tp4712934p4713396.html
Maybe itâs still possible to find it in the git history of the current pharovm repository?
Cheers, Henry
On Wednesday, October 13, 2021, Sven Van Caekenberghe <sven@stfx.eu> wrote: Hi Henry,
Thanks for your reply.
I created an issue: https://github.com/pharo-project/pharo/issues/10156
Can we still find the file 'sqExternalSemaphores.c' from the old emails ?
Sven
On 12 Oct 2021, at 19:20, Henrik Sperre Johansen <henrik.s.johansen@veloxit.no> wrote:
I believe Igor Stasenko made changes so signals could no longer get lost, but Iâm not sure it ever got integrated in Cogâ¦
https://forum.world.st/Issue-5965-in-pharo-Can-t-allocate-semaphores-Virtual...
Maybe worth revisiting/reviewing? If it works, the limitation would be gone, and the image warnings could be removed with no ill side-effects.
Cheers, Henry
On 12 Oct 2021, at 18:55, Henrik Sperre Johansen <henrik.s.johansen@veloxit.no> wrote:
Hi Sven - youâre correct the main intent was to make it clear one should not really be setting this from the image during runtime - and make it as obvious as possible when it had.
The way I remember it, the Error was only raised if the image is in development mode - so a developer running into this, would have the limit raised, but also get a debugger (which is hard to ignore) explaining the issue, while in runtime mode, the SystemNotification would get handled by logging so it would at least be possible to spot that it had happened.
If thereâs a less confusing way to achieve that it would be a good thing, but Iâm not sure removing all warnings would be the best way, as the related i/o problems would be nearly impossible to determine the root cause of.
Why forked? No idea/canât rememberâ¦
Cheers, Henry
On 12 Oct 2021, at 16:01, Sven Van Caekenberghe <sven@stfx.eu> wrote:
Hi,
The ExternalSemaphoreTable is an important resource since it holds 3 Semaphores per Socket (network connection). The size of ExternalSemaphoreTable determines how many such resources can be used concurrently.
VM parameter 49 holds this limit, it is accessed by VirtualMachine>>#maxExternalSemaphores. The default for Pharo 9 seems to be 256 currently.
There is a method VirtualMachine>>#maxExternalSemaphoresSilently: that can be used to set this limit higher. This seems to work.
Now, there are some limitations here: the number must be a power of 2, higher that the one set before and lower than 64K. BTW the number of file descriptors for files and sockets is also limited by the OS for its processes.
Also, while growing the table, the Semaphores become unavailable to the VM for a very short period of time, so they could miss signals, leading to IO issues.
For this reason, it seems that the design goal was to not allow this to grow during normal use, only during startup or upfront configuration.
Now comes the weird thing: the implementation/code of VirtualMachine>>#maxExternalSemaphores:
This is the method called when the current table is too small and when it wants to grow larger than the current VirtualMachine>>#maxExternalSemaphores.
The implementation calls #maxExternalSemaphoresSilently: in a forked thread, after signalling a SystemNotification with a full explanation. But why should this be forked in the first place ?
Then in the main thread, it signals an error ! Presumably because the design goals mentioned before.
These two actions are in conflict.
All this is probably the result of well meant evolution and refactoring, but it is quite confusing.
My first idea would be: let the caller raise the error and don't try raising the limit, remove the current implementation and replace it by what is now #maxExternalSemaphoresSilently:
What do you think ?
Sven