-
Notifications
You must be signed in to change notification settings - Fork 26
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Crash using medium test case #263
Comments
I have been seeing this type of problem in my tests also. It's a |
Hi @ericleigh007 - thank you for reaching out. |
I published this on github here: I did upgrade to the latest Netherite extension before trying again, and I immediately got the Out of Memory exception. I now wonder if things will be different if I follow any steps, such as redeploying my event hub. I've already used the included "Cleanup Function" in order to re-create the storage account, and that didn't have an effect. Unlike some of the other problems I'd reported before with this same benchmark, this one appears to be a show-stopper. When it occurs, the entire benchmark fails, with no retries -- nothing, and for some reason, it is happening much more often. Any workaround I should try, that you've discovered? |
Latest crash, in context, showing the orchestrators beginning to launch, but encountering this issue.
|
After completely removing all of the taskhub containers from the storage account, I have not experienced this error yet. |
Glad to hear you've found a workaround, @ericleigh007. |
I am seeing OutOfMemory problems in some of my tests also but am having a very hard time pinning down the reason, so I can't really tell if it is the same reason or not. |
@ericleigh007: I'm getting a 404 for this link: https://github.com/ericleigh007/DurableFunctionBenchmark-Ne/tree/master |
Big apologies there -- public now. |
I found an issue that was introduced with the new Faster parameter settings introduced in 1.3.3 that caused an excessive amount of memory to be allocated and could be responsible for a lot of recent issues caused by I also found several other issues that can cause memory problems which are a bit harder to fix and I will address those in additional work. Also, just out of curiosity, are you purging or replacing old orchestration instances? Having the total number of orchestration or entity instances in storage continuously grow can create memory issues with Netherite also. You can use purge timer functions as described here: #229 (comment) |
Hi Sir, |
FYI, we just released 1.3.5 which has several fixes that could potentially help with this situation. |
@sebastianburckhardt , we're using 1.3.5 and still seeing lots of OutOfMemory issues like this one
we're using basic plan, 1 unit and 12 partitions. This error happens to different partition numbers, not just this one. i've seen in 2, 5, 7, 9, etc the azure functions app is on consumption linux plan we're running tests with a relatively low amount of data being sent. about 10k messages with a few bytes each |
The latest release 1.4.0 addresses this issue. It contains a blob-batching optimization #275 that improves the performance for cases where partitions or clients transmit medium to large amounts of data (in terms of either total size, or number of messages). It also adds better handling for Out-Of-Memory issues. I think there is a good chance it will fix the issues you observed. If not, let me know and I will take another look. |
New test case testing load with 20 orchestrators with 20 activities each encountered a crash:
The text was updated successfully, but these errors were encountered: