Server crashes semi-regularly
Running a modded Minecraft server and every once in a while, a tick takes longer than 60 seconds and the server crashes. I have uploaded the crash report for the latest of these occurrences. The server has DH and the crash report does mention a far away chunk but beyond that, I'm not sure what exactly to look for to hunt down the cause. Any assistance appreciated!
35 Replies
crash-2025-07-12_21.10.25-server.txt uploaded to mclo.gsUploaded by thebestmaybe1101
you can disable watchdog by setting
max-tick-time in the server.properties to -1Wouldn't that just kind of be ignoring the problem?
worked on my server :shrug:
never caused any issues
Something's still slowing down ticks, it's just that the server will stop doing anything about it lol
Ig I can do that as a temporary fix at the very least
Are there smaller lag spikes that don't reach 60 sec
If so try spark
Will def check that out
Have a look at this part https://spark.lucko.me/docs/guides/Finding-lag-spikes#:~:text=The%20%2D%2Donly%2Dticks%2Dover,laggy%20ticks%20to%20worry%20about.
Finding the cause of lag spikes | spark docs
Lag spikes occur when a small number of ticks (or sometimes just one tick) takes a long time to execute.
Ok the server seems to have stalled or something. I assume this might be the same problem as yesterday, but now
max-tick-time has been set to -1 so the server has been allowed to play this problem out instead of crashing due to a tick taking too longmessage.txt uploaded to mclo.gsUploaded by thebestmaybe1101
Here's a fragment of the logs
did it ever unfreeze? a spark log would be very helpful if so
Nope
I've long since timed out but the server thinks I'm still online
It seems that the mod connecting the server to a Discord channel still works - messages sent in Discord are being relayed to the logs
There is a mod called Fast Backups which uses Git to backup the world periodically, looking at the errors, maybe that's what's freaking out
earlier in the logs ^
Backup starting and the mountain of errors are attributed to the same thread in the logs
I'll try asking on the Discord for Fast Backups
Hmm, backing up the server using Git in an online state?
How big is your World or how much is pregenerate? I have a similar issue without Fast Backups. In my case I think DH didn´t release my RAM and crash after a period of time. Maybe your problem is different but it seems pretty nearby.
How long is the time difference between each crash. In my case with a really big world it is after 1-2 days.
I pregenerated a 256 chunk radius in all 3 dimensions.
The time difference is inconsistent.
There were several crashes throughout day 1. Then I changed he max-tick-time configuration.
It was stable for a full day and froze at the end of day 2. Restarted it, froze again after about 1.5 hours.
It's the server's 3rd day now. Been up for about 5 hours now without Fast Backups.
The errors seem Fast Backups related to me but it could also be that they are a symptom, rather than a cause
I would not trust a solution that uses Git for backups on an active file, instead, something like Rsync would be better. With Rsync, it's specially designed to operate on active files.
Not a suggestion, but something of note. There's no mod, to my knowledge, that uses Rsync tech for backups.
declare variables
BCKHOME=/home/backup
DAY=$BCKHOME/day
HOUR=$BCKHOME/hour
MCHOME=/usr/local/minecraft/Fabric
WORLDPATH=$MCHOME
DATE=
date +%Y%m%d-%H%M%S
BCK=$HOUR/$DATE
LOG=$BCKHOME/backup.log
TGZLOG=$BCKHOME/backup_tgz.log
DELDATE=date -d "-1 week" +"%Y%m%d"
#DEL=$HOUR/$DELDATE
main
echo "---------------" >> $LOG
echo "sync from "$WORLDPATH" to "$BCKHOME >> $LOG
echo "/bin/sh/rsync "$WORLDPATH" to "$BCK >> $LOG
echo $DATE >> $LOG
/usr/bin/rsync -av $WORLDPATH $BCK >> $LOG
echo "---------------" >> $LOG
/bin/tar -zcvf $BCK.tgz $BCK >> $TGZLOG
/bin/rm -rf $BCK
This is my Backup skript im using Debian and use these scripts in my crontab
I use a similar scrip0t to delete the backups aftzer a period of timeThat's, uh, overkill...
I use day hour and week backups. because of space issues. I tar and zip my backups to minimize the impact.
sorry because of the complicaterd script but it worked fine for maybe 7 years or so
Finally got a response from someone on Fast Backup server, it might be a problem with saving Distant Horizons files while they're being written to
Folks, I'm back to square one 😭
The server froze again
Before the freeze was preceded by a bunch of errors that seemed related to Fast Backup. Now it's been running without Fast Backup and it froze anyway. This time, no errors in logs.
None leading up to it that is
message.txt uploaded to mclo.gsUploaded by thebestmaybe1101
Logs leading up to it
Timed out waiting for world statistics error from the thread used by Spark begins at one point. From there on out, it's that error every minute with Error sending packet clientbound/minecraft:disconnect sprinkled inThe LOD file is a single huge file and cannot be made a copy of while the file is actively being written to. I can see why that would trigger an error with Fast Backup.
CPU utilization shot up to 100% at the time of freeze and then flatlined
At this point I'm not even sure where to look for a cause. Spark is of no use by the time the server is frozen
Hmm, can't seem to make any sense with this log. There's not enough discerning information to go after.
Exactly
It just kinda stops at one point
For now I'm gonna be setting max-tick-rate back to normal and modifying the start script to auto-restart on crash
I don't know if this is luck or actual progress but this uptime is a record so far
The thing that's different this time: I removed Distant Horizons