Random TPS drops

Discussion in 'Performance Tweaking' started by BiGUNMAN, Jul 8, 2015.

  1. So I've attempted to fix this multiple ways, modified spigot.yml and bukkit.yml. I've referred to: https://www.spigotmc.org/threads/guide-optimizing-spigot-now-covering-paperspigot.21726 for fixing, haven't tried PaperSpigot yet, but might end up doing so if necessary.

    Very randomly I've been getting many lag spikes, and just today the server decided to stay at an average of 10-15 tps, when before it was pretty decently operating at 20tps.

    Spigot.yml: http://pastebin.com/77enFD45
    Bukkit.yml: http://pastebin.com/h8eZB71T

    I have the server restart every 12 hours, and the timings are all after a manual restart I did 30 minutes before, in attempt to fix the lag which started to occur.

    Timings from today:
    (Probably way to many)

    Unnecessary info (I think): http://gyazo.com/d8c470ddbdbfa6d6de9cedab23359d7a
    Yey... : http://gyazo.com/d70a6d033cbc50b532d49a2201d1d0ac

    Regarding server I'm running a OVH E5 1620V2, and SSDs - The only thing running is a bungeecord and this 1 server instance.

    Its never lagged this bad before, only minor lag spikes here and there every now and then. So.. Yeah. Help would be greatly appreciated.
    #1 BiGUNMAN, Jul 8, 2015
    Last edited: Sep 7, 2015
  2. Please make new timings and have them longer then 2 seconds. The longer the better.
  3. Apart from districts whichs cause most of your lag you also have problems with redstone (doTick) and chunk loading (do you use shared hosting? seems like your disk is pretty slow).
  4. As stated above I have an OVH dedi with SSDs (can give more details if necessary regarding that). So yeah :/
  5. I overread that, sorry. Maybe you should check your SSDs SMART status or make a speedtest to get more informations about them. But you should start with searching large redstone clocks to reduce doTickPending.
  6. I believe this is the SMART status although I have no idea what the data means: http://gyazo.com/addaba6a7121da4aba65287b1a7159ad (I have intel 730s in softraid x3). When looking at OVH's "real time monitoring" for CPU, although I don't know how accurate that is, it never has really gotten over 20% load.

    Have no idea how I would just randomly fly around my world looking for large redstone clocks :/
    Unless a plugin solution can also work: https://www.spigotmc.org/resources/redstone-clock-preventer-2.6678/
  7. If you have WorldGuard you could try /stoplag This will disable redstone, tnt explosions, water/lava flow and remove entities (as well as other things). See if that makes the server run any better.
  8. I need those functionalities working on my server so it wouldn't really help at the end of the day, but it most definitely will improve performance.
  9. You could try to see if it makes a difference but there is no guarantee that it will fix the problem but its worth a try. As for the crash I'm not too sure what caused it to crash. See if there is a new update available for 1.8.7. I also saw "world edit" a couple of times throughout the crash so make sure that its updated but i don't think its the cause.
  10. So I got this random spike today: http://timings.aikar.co/?url=12308166 - All I can tell is that "MorphGUI", "HiveJumpPads" and "Districts" are making a decent hit, but idk what else might be effecting it.