FOSS gamedev and creative worlds

Since upgrading our server to 0.4.6 we are experiencing the following
new problem:

Sometimes when placing a block it disappears immediately. Trying to
place it again at the same spot often fails over and over again, but
eventually will work. But sometimes some waiting and running around
is needed before placing the block in such an peculiar place succeeds.
Placing blogs at other places in the meantime usually works.

The effect reminds me remotely of problems caused by server lacks (the
immediate disappearing of placed blocks) but is obviously different.
(The problem only occurs when trying to place a block on some specific
places, and persists longer than the usual lagging).

Unfortunately I wasn't able to find a way to reproduce the problem at
will.

This is a regression, with 0.4.5 on the same server there were no such
problems...

minetest server and client build from latest git on Debian Squeeze.

This is a private mt server with less then 10 users, accessed remotely
via internet. The described problem even appears with only one active
user (no other users logged in).

Has anyone experienced similar problems?

I've seen this happen for me as well. No error messages or warnings of any kind - the block just disappears inexplicably, at least as recently as git from a couple of days ago. Not sure if it still happens in current HEAD.

Does the message "[playername] tried to place node in unloaded position [pos]" appear in debug.txt with debug_log_level = 4 in minetest.conf?

Ah, digging through the logs explains what was going on in my case. Swilde's problem is likely different from mine (a simple lack of 'interact' privilege, which the server didn't bother to warn me about).

That is the fix to the replace bug. Either that, or you successfully place the block by replacing one you didn't intend to :P

PilzAdam wrote:Does the message "[playername] tried to place node in unloaded position [pos]" appear in debug.txt with debug_log_level = 4 in minetest.conf?

I'll increase debugging to check, but I doubt this is the cause, as the problem appears in the middle of visible (and therefor I assume loaded) terrain.

mauvebic wrote:That is the fix to the replace bug. Either that, or you successfully place the block by replacing one you didn't intend to :P

Sorry, but I dont fully understand what you wrote. Are you saying the problem I'm seeing is the result of an fix for another problem?

I just saw:

commit 6e4fdf37baaeaf1048210e5519822ebd68dd413b
Author: sapier <Sapier at GMX dot net>
Date: Tue Apr 9 23:16:13 2013 +0200

fix objects colliding with its own collision boxes

Maybe this fixes my problem? I'll give it a try...

swilde wrote:

PilzAdam wrote:Does the message "[playername] tried to place node in unloaded position [pos]" appear in debug.txt with debug_log_level = 4 in minetest.conf?

I'll increase debugging to check, but I doubt this is the cause, as the problem appears in the middle of visible (and therefor I assume loaded) terrain.

We had a node replacement bug where the pointed node was replaced. The bug is fixed by not replacing the node and writing this message to the log. It also happened in loaded areas (because some strange status of the EmergeManager).

swilde wrote:I just saw:

commit 6e4fdf37baaeaf1048210e5519822ebd68dd413b
Author: sapier <Sapier at GMX dot net>
Date: Tue Apr 9 23:16:13 2013 +0200

fix objects colliding with its own collision boxes

Maybe this fixes my problem? I'll give it a try...

That doesnt fix it. It has nothing to do with the problem.

PilzAdam wrote:Does the message "[playername] tried to place node in unloaded position [pos]" appear in debug.txt with debug_log_level = 4 in minetest.conf?

swilde wrote:I just saw:

commit 6e4fdf37baaeaf1048210e5519822ebd68dd413b
Author: sapier <Sapier at GMX dot net>
Date: Tue Apr 9 23:16:13 2013 +0200

fix objects colliding with its own collision boxes

Maybe this fixes my problem? I'll give it a try...

That doesnt fix it. It has nothing to do with the problem.

While I trust you that this fix is unrelated to the problem you had in
mind, the problem I reported here disappeard after the last update.

So this seems to be solved for us for now. Thanks to every body for
the support! :)

I just got problems removing and placing blocks again. This time they are quite general (can't access any blocks) and I get messages in the log, like this:

00:05:26: ACTION[ServerThread]: Player swilde tried to access [node under=-30905,14,-14167 above=-30904,14,-14167] from too far: d=339047, max_d=140. ignoring.

Another player on our server had similar problems yesterday...

swilde wrote:I just got problems removing and placing blocks again. This time they are quite general (can't access any blocks) and I get messages in the log, like this:

00:05:26: ACTION[ServerThread]: Player swilde tried to access [node under=-30905,14,-14167 above=-30904,14,-14167] from too far: d=339047, max_d=140. ignoring.

Another player on our server had similar problems yesterday...

If you are sure that they dont use hacked clients then try

in minetest.conf

PilzAdam wrote:

swilde wrote:I just got problems removing and placing blocks again. This time they are quite general (can't access any blocks) and I get messages in the log, like this:

00:05:26: ACTION[ServerThread]: Player swilde tried to access [node under=-30905,14,-14167 above=-30904,14,-14167] from too far: d=339047, max_d=140. ignoring.

Another player on our server had similar problems yesterday...

If you are sure that they dont use hacked clients then try

Your phone or window isn't wide enough to display the code box. If it's a phone, try rotating it to landscape mode.

Code: Select all

disable_anticheat = true

in minetest.conf

As I trust them, so I'll try it -- but i suspect there is some seriouse bug at work here.

1. I never saw that problem in plder versions

2. When this problem happend and the player logs out and back in, he beamed to a different position, somewhere else on
the world. And that new position seems to match the old position of _another_ player...

So it seems, somehow the server confuses the players. Is that possible? Any ideas?

swilde wrote:

PilzAdam wrote:

swilde wrote:I just got problems removing and placing blocks again. This time they are quite general (can't access any blocks) and I get messages in the log, like this:

00:05:26: ACTION[ServerThread]: Player swilde tried to access [node under=-30905,14,-14167 above=-30904,14,-14167] from too far: d=339047, max_d=140. ignoring.

Another player on our server had similar problems yesterday...

If you are sure that they dont use hacked clients then try

Your phone or window isn't wide enough to display the code box. If it's a phone, try rotating it to landscape mode.

Code: Select all

disable_anticheat = true

in minetest.conf

As I trust them, so I'll try it -- but i suspect there is some seriouse bug at work here.

1. I never saw that problem in plder versions

2. When this problem happend and the player logs out and back in, he beamed to a different position, somewhere else on
the world. And that new position seems to match the old position of _another_ player...

So it seems, somehow the server confuses the players. Is that possible? Any ideas?

Do you use the carts mod or any other mod were the player is attached to an object?

PilzAdam wrote:Do you use the carts mod or any other mod were the player is attached to an object?

No.

For the first time for a looong time the original Problem reappeared -- and yes, PilzAdam you were right, the log shows:
16:47:52: INFO[ServerThread]: swilde tried to place node in unloaded position (-30912,16,-14632)

Do I understand correctly that there is no fix for this yet?

Anybody know , how detect disappearing of blocks in debug.txt with option "debug_log_level = 4" in minetest.conf ?

This is the fix for it, I think... https://github.com/minetest/minetest/pull/1153