First of all, thank you for taking the time to read this. I'm at my wit's end on this one.
I set up a FTB Revelation server a few weeks ago, and everything seemed to be going fine for some time. However, we are now seeing 2-3s tick skips every 45-60s. TPS still reads at a smooth 20, but it has a noticeable impact on gameplay, as the server simply stops responding for two or three seconds every minute. This occurs constantly every 45-60s while at least one player is connected and about every 4-5 minutes while no players are connected. Outside of these really aggressive lag spikes, the server runs buttery-smooth.
As for the server itself, it's hosted on a cloud service with scalable 2-4 cores CPU, 4-8G RAM, and plenty of SSD space running in Ubuntu. I list this as scalable because switching between the lower and higher ends of the parameters does not affect this lag issue. It's currently set on the upper end. When monitoring with top, CPU usage occaisionally spikes to 100%, but memory averages a cool 33%, so it doesn't seem to need more RAM thrown at it.
I've searched all over the internet for people that are having the same issue, and I have found at least two. However, there was no solution found in those threads. Relevant links:
https://www.minecraftforge.net/foru...nciding-quotcant-keep-upquot-messages-on-smp/ (I am seeing the same world leakage messages in debug.log, but I've heard those are likely unrelated.)
https://forum.feed-the-beast.com/threads/ftb-revelation-server-sag-spikes-every-minute.293472/ (From what I can tell from this, it does seem likely that my problem is related to chunks getting unloaded, but I can't tell why that's happening or why there'd be a mass unload every minute. Using ae2 ChunkLogger, I do see this behavior, but isn't that normal? It happens constantly, even when there is no lag spike.)
Some things I have tried/additional information:
- I have installed LagGoggles and TickProfiler and can provide any of their output data, if it helps. Upon running around with LagGoggles multiple times, I don't find any entities in loaded regions that are causing a grievous amount of lag, so it really doesn't appear to be an issue with too many entities.
- We do have a few chunks perma-loaded, but it's so few and LagGoggles says they're low cost, anyway.
- We also have some more mods added that may be related (no log has indicated this, though.):
- compactmachines3-1.12.2-3.0.18-b278.jar
- EmbersRekindled-1.13-hotfix2.jar
- mysticallib-1.12.2-1.3.8.jar
- MysticWorld-1.12.2-1.4.7.jar
- OpenComputers-MC1.12.2-1.7.4.153.jar
- rftoolsdim-1.12-5.71.jar (possible culprit? We have two loaded dims with no anchors in them.)
- Roots-1.12.2-3.0.17.jar
- secretroomsmod-1.12.2-5.6.4.jar
- Soot-1.6.jar
- SoulShards-TOW-1.12-2.7.6-56.jar
- LagGoggles-FAT-1.12.2-4.9.jar
- I updated Flux Networks to the latest version that is compatible with 1.12.2.
Attached are a recent latest.log and debug.log with the start-up bits removed. (Appended .txt per forum rules)
If there's any information I can provide which I left out, please don't hesitate to ask. I'm seriously confused as to what's causing this, but would really like to have a solution. Your assistance is GREATLY appreciated.
I set up a FTB Revelation server a few weeks ago, and everything seemed to be going fine for some time. However, we are now seeing 2-3s tick skips every 45-60s. TPS still reads at a smooth 20, but it has a noticeable impact on gameplay, as the server simply stops responding for two or three seconds every minute. This occurs constantly every 45-60s while at least one player is connected and about every 4-5 minutes while no players are connected. Outside of these really aggressive lag spikes, the server runs buttery-smooth.
As for the server itself, it's hosted on a cloud service with scalable 2-4 cores CPU, 4-8G RAM, and plenty of SSD space running in Ubuntu. I list this as scalable because switching between the lower and higher ends of the parameters does not affect this lag issue. It's currently set on the upper end. When monitoring with top, CPU usage occaisionally spikes to 100%, but memory averages a cool 33%, so it doesn't seem to need more RAM thrown at it.
I've searched all over the internet for people that are having the same issue, and I have found at least two. However, there was no solution found in those threads. Relevant links:
https://www.minecraftforge.net/foru...nciding-quotcant-keep-upquot-messages-on-smp/ (I am seeing the same world leakage messages in debug.log, but I've heard those are likely unrelated.)
https://forum.feed-the-beast.com/threads/ftb-revelation-server-sag-spikes-every-minute.293472/ (From what I can tell from this, it does seem likely that my problem is related to chunks getting unloaded, but I can't tell why that's happening or why there'd be a mass unload every minute. Using ae2 ChunkLogger, I do see this behavior, but isn't that normal? It happens constantly, even when there is no lag spike.)
Some things I have tried/additional information:
- I have installed LagGoggles and TickProfiler and can provide any of their output data, if it helps. Upon running around with LagGoggles multiple times, I don't find any entities in loaded regions that are causing a grievous amount of lag, so it really doesn't appear to be an issue with too many entities.
- We do have a few chunks perma-loaded, but it's so few and LagGoggles says they're low cost, anyway.
- We also have some more mods added that may be related (no log has indicated this, though.):
- compactmachines3-1.12.2-3.0.18-b278.jar
- EmbersRekindled-1.13-hotfix2.jar
- mysticallib-1.12.2-1.3.8.jar
- MysticWorld-1.12.2-1.4.7.jar
- OpenComputers-MC1.12.2-1.7.4.153.jar
- rftoolsdim-1.12-5.71.jar (possible culprit? We have two loaded dims with no anchors in them.)
- Roots-1.12.2-3.0.17.jar
- secretroomsmod-1.12.2-5.6.4.jar
- Soot-1.6.jar
- SoulShards-TOW-1.12-2.7.6-56.jar
- LagGoggles-FAT-1.12.2-4.9.jar
- I updated Flux Networks to the latest version that is compatible with 1.12.2.
Attached are a recent latest.log and debug.log with the start-up bits removed. (Appended .txt per forum rules)
If there's any information I can provide which I left out, please don't hesitate to ask. I'm seriously confused as to what's causing this, but would really like to have a solution. Your assistance is GREATLY appreciated.