Server Keeps Crashing & Client Problems

Hi,

I’d just like to let you know about problems that we are having:-

1, The server keeps crashing. Sometimes the process is running but web page isn’t responding and other times the process isn’t running at all.
2. It seems that if the client starts an image backup and then the server crashes the client is no no-longer able to do any type of backup. Message appears “A backup is already running. Could not start another one” However the client Icon in the system tray is white.
3, A full file backup that has run on my clent was not logged on the server at all. A check of the files did reveal that the backup was completed.
4, I have set the settings to now use the server settings and not allow client to override. The clients don’t seem to have the settings as laid down by the server on their settings screen.

Backup Server is running on Ubuntu 10.10
Temp folder and Backup destination are on a on a NFS share on a nas box
Client Windows 7 x64 SP1

I’m going to delete the client, remove software and reinstall the client.

Just thought i’d let you know the problems we’re encountering

Regards
Robert

I’m sorry that you are experiencing so many problems. I hope you want to help me fix them.

Regarding

  1. Sounds pretty bad. Two things: Could you set the loglevel in /etc/default/urbackup_srv to debug and tell me what the last log messages before the crashes are?
    The abundance of weird crashes also suggests something fundamentally wrong. I didn’t test the debian package on ubuntu. I think I will try that now. You can also compile the server, that way we can be sure it’s not some kind of dependency problem.
  2. Okay. Sorry. Fixed that now. (for next version)
  3. Hmm the server writes all files first (which can take some time) and then saves the log. Maybe it crashed in between. Did you set the filter to info? Otherwise it only shows logs with warnings or errors in them.
  4. That’s a feature. If you have more then one backup server one could still use the client settings. I’ll have to change that in the future so that its more clear.

Hi Uroni,

Don’t be sorry, you’ve done a grand job! :)

1, Yes tried debug but the webpage requests get logged in there and there is nothing that looked out of place when we checked the log after the service stopped. I’ll set it back to debug and wait for the next crash. Would it not run at all if it was dependency?

2, Thanks

3, That’s very probable. Yes I changed filter to all. I don’t know if the log would be better written as you’re going along instead of after?

4, Would it be possible to have a “refresh settings from server” option on the client? How often does the client get the settings from the server or is it only pulled down once? IS there a file that we can delete? How do we know what settings are inforced on the client?

I’ll post up the debug entry when I next have a crash.

Uroni,

I didn’t have to wait long!

9/16/11 15:18:25: HTTP: Waking up…790192
09/16/11 15:18:25: Incoming data for client…
09/16/11 15:18:25: Parsing done… starting handling request_num: 0 790195
09/16/11 15:18:25: HTTP: Removing user790205
09/16/11 15:18:25: HTTP: Sleeping…790216
09/16/11 15:18:26: HTTP: New Connection incomming 791232 s: 5
09/16/11 15:18:26: HTTP: Waking up…791232
09/16/11 15:18:26: Incoming data for client…
09/16/11 15:18:26: Parsing done… starting handling request_num: 0 791235
09/16/11 15:18:26: HTTP: Removing user791246
09/16/11 15:18:26: HTTP: Sleeping…791256
09/16/11 15:18:27: HTTP: New Connection incomming 792338 s: 5
09/16/11 15:18:27: HTTP: Waking up…792339
09/16/11 15:18:27: Incoming data for client…
09/16/11 15:18:27: Parsing done… starting handling request_num: 0 792340
09/16/11 15:18:27: HTTP: Removing user792351
09/16/11 15:18:27: HTTP: Sleeping…792361
09/16/11 15:18:28: HTTP: New Connection incomming 793365 s: 5
09/16/11 15:18:28: HTTP: Waking up…793366
09/16/11 15:18:28: Incoming data for client…
09/16/11 15:18:28: Parsing done… starting handling request_num: 0 793367
09/16/11 15:18:28: HTTP: Removing user793377
09/16/11 15:18:28: HTTP: Sleeping…793388
09/16/11 15:18:29: HTTP: New Connection incomming 794393 s: 5
09/16/11 15:18:29: HTTP: Waking up…794393
09/16/11 15:18:29: Incoming data for client…
09/16/11 15:18:29: Parsing done… starting handling request_num: 0 794394
09/16/11 15:18:29: HTTP: Removing user794404
09/16/11 15:18:29: HTTP: Sleeping…794415
09/16/11 15:18:29: HTTP: New Connection incomming 794613 s: 5
09/16/11 15:18:29: HTTP: Waking up…794613
09/16/11 15:18:29: Incoming data for client…
09/16/11 15:18:29: Parsing done… starting handling request_num: 0 794615
09/16/11 15:18:29: HTTP: Removing user794626
09/16/11 15:18:29: HTTP: Sleeping…794636
09/16/11 15:18:31: xxx-xxx: Connecting to client…
09/16/11 15:18:31: xxx-xxx: Loading filelist…
09/16/11 15:18:32: xxx-xxx Starting incremental backup…
09/16/11 15:18:32: HTTP: New Connection incomming 797338 s: 20
09/16/11 15:18:32: HTTP: Waking up…797338
09/16/11 15:18:32: Incoming data for client…
09/16/11 15:18:32: Parsing done… starting handling request_num: 0 797339
09/16/11 15:18:32: xxx-xxx: Calculating file tree differences…
09/16/11 15:18:32: HTTP: Removing user797907
09/16/11 15:18:32: HTTP: Sleeping…797917

Okay. Thanks. Looks like the crash is happening while calculating file tree differences. Could you verify that by doing a full backup and see if that works and/or if it crashes all the time after “Calculating file tree differences…”?

If the filenames on the client are not sensitive could you send me your /var/urbackup/clientlist_*.ub from the server and C:\Program Files\UrBackup\urbackup\data\filelist.ub after such a crash happend?(to martin@urbackup.org) That way I could try to reproduce the crash.
The other option would be building a debug build (make CXXFLAGS=-g) and running the server in gdb and sending me a stacktrace after the crash.