0.4.0 beta 2 fills up cache-directory (after restart?) with 1647 files / 877 MB

[sander]

This has happened a few times: SABnzbd 0.4.0 beta 2 filling up the cache directory with 1647 files with a total size of 877MB:

sander@ubuntu804:~/SABnzbd-0.4.0Beta2$ du -BM /home/sander/cache/
877M    /home/sander/cache/
sander@ubuntu804:~/SABnzbd-0.4.0Beta2$ ll /home/sander/cache/ | wc -l
1647
sander@ubuntu804:~/SABnzbd-0.4.0Beta2$

I believe it happens after an ungraceful stop and/or restart.

SABnzbd did not recover from this and my disk got full.

[nzb_leecher]

what if you set a max to cache? I set mine to 350M, had no problems yet after many restarts

[sander]

The cache (.sabnzbd/cache/) filled up again until my disk was full.

I will try the cache limit.



Traceback (most recent call last):
  File "/usr/lib/python2.5/logging/__init__.py", line 753, in emit
    self.flush()
  File "/usr/lib/python2.5/logging/__init__.py", line 731, in flush
    self.stream.flush()
IOError: [Errno 28] No space left on device
2008-04-18 23:16:58,147::WARNING::[assembler] missing
Traceback (most recent call last):
  File "/usr/lib/python2.5/logging/__init__.py", line 753, in emit
    self.flush()
  File "/usr/lib/python2.5/logging/__init__.py", line 731, in flush
    self.stream.flush()
IOError: [Errno 28] No space left on device
2008-04-18 23:16:58,186::WARNING::[assembler] missing
Traceback (most recent call last):
  File "/usr/lib/python2.5/logging/__init__.py", line 753, in emit
    self.flush()
  File "/usr/lib/python2.5/logging/__init__.py", line 731, in flush
    self.stream.flush()
IOError: [Errno 28] No space left on device
2008-04-18 23:16:58,222::WARNING::[assembler] missing
Traceback (most recent call last):
  File "/usr/lib/python2.5/logging/__init__.py", line 753, in emit
    self.flush()
  File "/usr/lib/python2.5/logging/__init__.py", line 731, in flush
    self.stream.flush()
IOError: [Errno 28] No space left on device

[sander]

What do I fill out here? 300? 300000? 300MB?

"Article Cache Limit:
Cache articles in memory (reduce disk access)."

[shypike]

The cache directory and the memory cache are two completely different things.
The memory cache is for keeping articles in memory instead of writing them to disk.
Setting it high enough so that SABnzbd can hold a complete file, reduces disk access.
I'd say: set it to 70M if you can spare the memory.

Another tip is to enable the disk space check for the download dir.
It's better to pause SABnzbd before the disk fill up.

These tips will not solve your problem with the bad resume after a crash.
When SABnzbd is bluntly terminated, you never know what it will do on restart.
Although I'm surprised and puzzled by what happens on your system.
Maybe some consistency checks can be added to prevent too wild behaviour.

[sander]

The second (third), fresh try of a download failed too; with 300 MB to go, SABnzbd had stopped.

Can I send the NZB to a mail adresses for inspection? What's the address?

[shypike]

bugs at sabnzbd.org

[sander]

Even with SAB 0.3.4, SAB stops with still 30 MB or so to go. The directory is still in "incomplete". No (!) output on stdout / stderr.

I've sent the NZB to the bugs-address

And I've created a new NZB from another search engine.

[sander]

Some result with a NZB from another search engine: SAB stops with still 60 MB to go.

Strange.

BTW: It doesn't have to be the NZB; it could also be something on my system.