Server Discussion New update (Furious Koala) Server Issues

It appears that the server starts lagging off and crashes eventually at random times. We do not know how many players are online because -sigh- there's no proper playerlist yet. Although we had put a player-cap of 50 in place, there were no signs of increasing lag, just a random full stop and crash.

Owner says there's an issue related to the NetSocket protocol, causing exceptions and leading to a crash, he thinks that it is being caused by outdated clients trying to connect and failing, taking the server down as well.

I assume a hotfix is on the way for this? Because I'm fairly sure we're not the only ones with this issue.

EDIT:
This is pretty important information

So that's the problem right there, devs. It connects with everything that has been found out about it previously, can we expect a fix for this?

 

Why is this a post?

 

Because I am re-wording it......................

 

Gotcha!

 

-Just a thought here:
Did your new update, get the required permissions, within its environment? New/Updated folders don't always receive the hierarchy of rules from their parent folders. It looks like, from first glance, that the issue is a read/write issue, and the 'liquid' could not be created, but doesn't stop the server from trudging on. Since the liquid is called so often, it casts the appearance that the error maybe elsewhere.

Can you confirm that the folder permissions have not changed?

 

The issue is within the server software, not within our setup..

People simply hang randomly, not being able to chat/break things, and then they have to reconnect to be able to play again.

EDIT: Actually, the issue may be within the clients, some lag+crash and others don't..

EDIT: confirmed that it happens when certain clients connect, we suspect outdated ones?

 

I'm having the same issue. It crashes like every hour.

 

It is definitely caused by outdated clients connecting, or something else that some clients seem to share. The issue is becoming less and less frequent as well, would be logical because more people are updating of course, and more parts of the worlds are generated then

 

I believe the Steam rollout goes by timezone. I am in Chicago, IL USA, and my update started at 72am local time. So maybe by this evening( GMT ) it will work itself out?

 

Nah, this update was rolled out worldwide.

 

This update Furious is a nightmare, my restart script as never restarted so much the server Seems always a
Segfault here I get

Centos 6.5 64-bit
Furious Koala 635

as log I have (pm me for full log it has lots of private ips)

Code:

Warn: Perf: UniverseServer::handleQueuedConnections.one millis: 40000
Warn: Perf: UniverseServer::handleQueuedConnections millis: 40000
Info: UniverseServer: Kicking client 'john' <6> () due to inactivity
Info: UniverseServer: Shutting down ship world for client 'john' <6> ()
Info: NetSocket write loop shut down: Socket closed
Info: UniverseServer: Client 'john' <6> () disconnected
Info: UniverseServer: Reaping client 'john' <6> () connection
Info: UniverseServer: Kicking client 'Hebi' <9> () due to inactivity
Info: UniverseServer: Shutting down ship world for client 'Hebi' <9> ()
Info: UniverseServer: Client 'Hebi' <9> () disconnected
Info: UniverseServer: Reaping client <24> () connection
Error: Segfault Encountered!
Error: Stack Trace...
./starbound_server(_ZN4Star15innerPrintStackEv+0x2c) [0xdc75bc]
./starbound_server(_ZN4Star10fatalErrorEPKcb+0x1d) [0xdc683d]
/lib64/libc.so.6(+0x329a0) [0x7f5ddf31e9a0]
./starbound_server(_ZNSt5dequeISt4pairIjN4Star17WarpCommandPacketEESaIS3_EE12emplace_backIIS3_EEEvDpOT_+0xc) [0x84737c]
./starbound_server(_ZN4Star14UniverseServer20ClientSocketCallback14packetReceivedERKSt10shared_ptrINS_6PacketEE+0xb9) [0x83dff9]
./starbound_server(_ZN4Star9NetSocket9putPacketERKSt10shared_ptrINS_6PacketEE+0x29) [0xa6e379]
./starbound_server(_ZN4Star9NetSocket10readerLoopEv+0x189) [0xa6e629]
./starbound_server(_ZN4Star10ThreadImpl9runThreadEPv+0x21) [0xdcc351]
/lib64/libpthread.so.0(+0x79d1) [0x7f5de002e9d1]
/lib64/libc.so.6(clone+0x6d) [0x7f5ddf3d4b6d]

 

I can't start the server it seems. Simply typing: "sudo service starbound start " doesn't do a thing. Is there a new way of starting the server after the update?

 

No.

 

Warn: I/O error from cens:55343 in NetSocket read loop or device closed: NetworkException: tcp recv error: Resource temporarily unavailable
./starbound_server(_ZN4Star13StarExceptionC2ERKSs+0xfe) [0xdc5c7e]
./starbound_server(_ZN4Star16NetworkExceptionC2ERKSs+0x11) [0xd75811]
./starbound_server(_ZN4Star9TcpSocket4readEPcm+0x155) [0xd749a5]
./starbound_server(_ZN4Star10DataStream8readDataEPcm+0x1e) [0xd3713e]
./starbound_server(_ZN4Star10DataStreamrsERh+0x16) [0xd37466]
./starbound_server(_ZN4Star9NetSocket10readerLoopEv+0xf8) [0xa6e598]
./starbound_server(_ZN4Star10ThreadImpl9runThreadEPv+0x21) [0xdcc351]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x7e0e) [0x7fb5113f8e0e]
/lib/x86_64-linux-gnu/libc.so.6(clone+0x6d) [0x7fb5107130fd]

Info: UniverseServer: Client 'cens' <2> (cens:55343) disconnected
Error: Segfault Encountered!
Error: Stack Trace...
./starbound_server(_ZN4Star15innerPrintStackEv+0x2c) [0xdc75bc]
./starbound_server(_ZN4Star10fatalErrorEPKcb+0x1d) [0xdc683d]
/lib/x86_64-linux-gnu/libc.so.6(+0x35250) [0x7fb51065f250]
./starbound_server(_ZNSt5dequeIjSaIjEE12emplace_backIIjEEEvDpOT_+0) [0x847070]
./starbound_server(_ZN4Star14UniverseServer20ClientSocketCallback6closedEv+0x57) [0x835867]
./starbound_server(_ZN4Star9NetSocket10readerLoopEv+0x318) [0xa6e7b8]
./starbound_server(_ZN4Star10ThreadImpl9runThreadEPv+0x21) [0xdcc351]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x7e0e) [0x7fb5113f8e0e]
/lib/x86_64-linux-gnu/libc.so.6(clone+0x6d) [0x7fb5107130fd]

Aborted

 

Are these new questions, or questions now pertaining to class101's CentOS thread?

At first glance, it seems to me, you need the new LUA and asset pack readers, as well as making sure that the OWNER/GROUP of the new files folders, didn't change.

 

The issue is nearly gone, devs better check this next time they are going to update... See what happens when older clients try to connect to an updated server. Another bit of proof is that the serverlists temporarily disabled their status checker for the updated servers because their query every 5 mins would make the server hang for a while, so they quickly removed it for the updated servers. That's why you see outdated playercap info and "NA/(number)".

 

Well its not nearly gone for me, I'm getting occasionally outdated players join, crashing the server.

Code:

Warn: Recovering SimpleDatabase at device: ../universe/alpha_42247126_7134209_242679_9_5.world
Warn: Recovery finished
Info: UniverseServer: Client 'Lynda' <43> (84.57.156.222:50314) connected
Warn: Perf: UniverseServer::handleQueuedConnections.one millis: 1901
Warn: Perf: UniverseServer::handleQueuedConnections millis: 1901
Error: Segfault Encountered!
Error: Stack Trace...
./starbound_server(_ZN4Star15innerPrintStackEv+0x2c) [0xdc75bc]
./starbound_server(_ZN4Star10fatalErrorEPKcb+0x1d) [0xdc683d]
/lib/x86_64-linux-gnu/libc.so.6(+0x35250) [0x7ffe93781250]
./starbound_server(_ZNSt5dequeISt4pairIjN4Star17WarpCommandPacketEESaIS3_EE12emplace_backIIS3_EEEvDpOT_+0xc) [0x84737c]
./starbound_server(_ZN4Star14UniverseServer20ClientSocketCallback14packetReceivedERKSt10shared_ptrINS_6PacketEE+0xb9) [0x83dff9]
./starbound_server(_ZN4Star9NetSocket9putPacketERKSt10shared_ptrINS_6PacketEE+0x29) [0xa6e379]
./starbound_server(_ZN4Star9NetSocket10readerLoopEv+0x189) [0xa6e629]
./starbound_server(_ZN4Star10ThreadImpl9runThreadEPv+0x21) [0xdcc351]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x7e0e) [0x7ffe9451ae0e]
/lib/x86_64-linux-gnu/libc.so.6(clone+0x6d) [0x7ffe938350fd]

Aborted

 

Ye still getting it as well now, it happens every 10-15 minutes or so, and the server simply closes then.

EDIT: Actually, it can happen within 5 minutes as well, this is pretty damn frustrating............

 

Liquid errors are nothing new (for me at least) they still spam my console as they always have.

 

Yea this is not about the liquid errors it seems, it's more about the fact that old clients can crash updated servers on join.

I'll remove it from the main post.