serversnapshotdeploy -keepfiles crashes the server and leaves the virtual server unrecoverable
Version: TeamSpeak Server 6.0.0-beta12.1 [Build: 1785239375], running in Docker
Host: Linux 6.18.33.2-microsoft-standard-WSL2 x86_64 (Docker Desktop), MariaDB 12.3.3 (database plugin “MariaDB plugin, version 3”)
Interface: SSH query (port 10022), logged in as serveradmin
License: none (“No License”, max virtualservers: 1)
Steps to reproduce
- Run a server with a single virtual server (sid 1). In the tries below it held a few channels and one or two connected clients.
- Optional: upload a file to a channel. Try 3 crashed with a file stored in the Default Channel.
- Send the following over the SSH query:
In try 3 the deploy command was sent exactly as the query documentation shows it, with the data passed on unchanged.use sid=1 serversnapshotcreate serversnapshotdeploy -keepfiles version=3 data=<data returned by serversnapshotcreate>
Expected
The snapshot is deployed. The files of channels that exist both on the server and in the snapshot are kept, as the documentation of -keepfiles describes.
Actual
serversnapshotdeploynever answers.- Within seconds the server process crashes:
In try 3 the container exited with code 139 (SIGSEGV).|VirtualServerBase|1 |stopped TeamSpeak server has crashed. A crashdump has been generated at "/var/tsserver/crashdumps/tsserver_6_0_0-beta12_1_linux_amd64_<timestamp>.dmp" - After the automatic restart, virtual server 1 cannot be loaded:
|ERROR |VirtualServer |1 |VIRTUALSERVER_DEFAULT_SERVER_GROUP points to 0 |ERROR |VirtualSvrMgr | |loading VirtualServer(1) failed to start, error: invalid group ID - From then on the virtual server is lost:
serverlist -allshows itoffline, with an emptyvirtualserver_name.use sid=1 -virtualfails witherror id=2560 msg=invalid group ID, and the server log addsstartServer() VirtualServer(1) failed to start, error: invalid group ID.serverdelete sid=1, also with-noevent, fails witherror id=1281 msg=database empty result set. This was tried after try 1 only.
- The only way to recover was wiping the database and setting the server up again.
- After try 1, while the virtual server hung in
deploy running, its WebQuery API key was rejected with5122 invalid apikey. This was not checked after tries 2 and 3.
Narrowing down
Every try created its own snapshot with serversnapshotcreate immediately before deploying it.
| Try | Restart right before | -keepfiles |
File in a channel | Result |
|---|---|---|---|---|
| 1 | yes | yes (after the parameters, with -mapping) |
no | hung in deploy running for over 20 minutes; after a container restart the virtual server could not start (same errors as above) |
| 2 | no | yes (after the parameters, with -mapping) |
no | server crashed, crashdump written, virtual server unrecoverable |
| 3 | no | yes (documented form: option first, no -mapping) |
yes, one CSV file | server crashed within seconds (exit code 139), virtual server unrecoverable |
| 4 | no | no (-mapping) |
not checked | ok in about 1 s |
| 5 | yes, deploy at uptime 0 | no (-mapping) |
yes, one CSV file | ok in 0.8 s; the file was no longer listed afterwards, as documented |
Before tries 2 and 3, a deploy without -keepfiles had already succeeded on the same server (tries 4 and 5).
Deploys without -keepfiles succeeded every time. The snapshot of try 1 was checked: it decompresses cleanly, has all sections, and its default groups point to groups it contains.
Log of try 3 (UTC)
The container exited with code 139 (SIGSEGV) and was restarted by Docker:
teamspeak6-01 | TeamSpeak server has crashed. A crashdump has been generated at "/var/tsserver/crashdumps/tsserver_6_0_0-beta12_1_linux_amd64_1789392034.dmp"
teamspeak6-db | 2026-09-14 13:20:37 15 [Warning] Aborted connection 15 to db: 'teamspeak' user: 'teamspeak' host: '172.20.0.3' (Got an error reading communication packets)
(the same warning for connections 6 to 14)
teamspeak6-01 exited with code 139 (restarting)
teamspeak6-01 | 2026-09-14 13:20:37.940245|INFO |ServerLibPriv | |TeamSpeak Server 6.0.0-beta12.1 (2026-07-28 11:49:35)
...
teamspeak6-01 | 2026-09-14 13:20:38.760266|ERROR |VirtualServer |1 |VIRTUALSERVER_DEFAULT_SERVER_GROUP points to 0
teamspeak6-01 | 2026-09-14 13:20:38.760370|ERROR |VirtualSvrMgr | |loading VirtualServer(1) failed to start, error: invalid group ID
teamspeak6-01 | 2026-09-14 13:20:38.763993|INFO |VirtualServerBase|1 |stopped
teamspeak6-01 | 2026-09-14 13:21:04.760524|ERROR |VirtualServer |1 |VIRTUALSERVER_DEFAULT_SERVER_GROUP points to 0
teamspeak6-01 | 2026-09-14 13:21:04.760604|ERROR |VirtualSvrMgr | |startServer() VirtualServer(1) failed to start, error: invalid group ID
teamspeak6-01 | 2026-09-14 13:21:04.765159|INFO |VirtualServerBase|1 |stopped
The last three lines come from use sid=1 -virtual.
Crash dumps
Two crash dumps are available:
tsserver_6_0_0-beta12_1_linux_amd64_1789390089.dmp(try 2)tsserver_6_0_0-beta12_1_linux_amd64_1789392034.dmp(try 3)
They are not attached because they may contain credentials. I can send them privately on request.
Note: This analysis was put together with the help of an AI assistant (Claude). It ran the query commands against my test server while developing a ServerQuery tool, and collected the results in one place. The server logs are copied from my own Docker output. Everything marked “not checked” was not verified.
