It seems that fullGC cannot be correctly identified.

[youngledo]

I have a lot of fullGC through the jstat command output (the same is true with GC Easy), but only once with jifa.

• jstat
  
  note: Don't pay attention to the number of times, because it is not the same time period.

• GC Easy
  For example, this time period is displayed 7 times:
  

I compared the time in gc.rog and found that this identifier is recognized as fullGC, but jifa doesn't seem to recognize it.

attachment

app-gc-pid6-2024-10-10_21-48-02.log

---

In addition, if the DingTalk QR code is invalid, can I still join the group?

[D-D-H]

JVM treats the two stop phase of cms concurrent gc as FGC, so the count of FGC increases by two when one CMS concurrent gc occurs.
However, Jifa doesn't treat the two phases as FGC.

[D-D-H]

In addition, if the DingTalk QR code is invalid, can I still join the group?

I'll update it.

[youngledo]

JVM treats the two stop phase of cms concurrent gc as FGC, so the count of FGC increases by two when one CMS concurrent gc occurs. However, Jifa doesn't treat the two phases as FGC.

However, this will be inconsistent with the FGC results output by jstat, which will cause confusion. This is consistent between jstat and gceasy.

It is recommended to add a description of this section.

[D-D-H]

Here is a good answer about this question: https://stackoverflow.com/questions/18306490/fullgc-happens-every-time-when-cms-work

It is recommended to add a description of this section.

Good suggestion.

[youngledo]

Here is a good answer about this question: https://stackoverflow.com/questions/18306490/fullgc-happens-every-time-when-cms-work

It is recommended to add a description of this section.

Good suggestion.

I saw it yesterday, but I feel RednaxelaFX great god is more convincing!

link:

• https://www.iteye.com/blog/rednaxelafx-1108768
• https://openjdk.org/groups/hotspot/docs/Serviceability.html#bjvmstat