polipo constantly clears out the cache

[Rolf Leggewie]

polipo keeps forgetting about the stuff it has cached, essentially rendering it useless as a caching proxy. I get "Discarded all objects, 10 + 2 objects left (1 chunks and 180 atoms used)." in the logs every few hours.

[Sean Young (techsy730)]

Sorry for the bump, but I think this behavior is intentional.

Polipo tries to write out entries from memory to disk every now and then until it hits the chunkLowMark in memory usage.
You can tune how aggressively it does so by playing with the optoins maxObjectsWhenIdle and maxWriteoutWhenIdle.
And of course, you can also make it flush less by rasing the chunkLowMark, so it is willing to hold more in RAM for long periods of time.

I highly reccomend checking polipo's info pages (use the command "info polipo" without the quotes) for a much better explaination, as well as much more information on many other options.

However, I do find it strange that is says "Discarded all objects". Either that is just a bad log message, or it is indeed flushing out the whole RAM cache too often. (I'm pretty certain that these discarded entries go to the disk cache when this happens)

[Rolf Leggewie (r0lf)]

No need to apologize for the bump, it's highly appreciated. But indeed it is as I wrote, the data is purged/discarded.

[Sean Young (techsy730)]

All of the RAM cache is dumped to disk spuriously?

Yea, that would be a bug.

[Sean Young (techsy730)]

Or is it not even reaching the disk?

(dang it, why can't I edit my comments)

[Rolf Leggewie (r0lf)]

FWIW, I am not experiencing this anymore with the git version. I guess this ticket can be closed when that is uploaded.

[Launchpad Janitor (janitor)]

This bug was fixed in the package polipo - 1.1.1-1

---------------
polipo (1.1.1-1) unstable; urgency=medium

  * release newest upstream version 1.1.1. Closes: #746750
    - use regerror to improve error reporting. Closes: #585766
    - correct return value from processObject in error case. LP: #988780
    - resume interrupted file download properly. Closes: #517479
    - vastly improve stability. LP: #1055418, LP: #988780, LP: #1168593
      Closes: #684742, Closes: #307801, Closes: #432561, Closes: #646709
    - handle non-existent host properly. Closes: #655851
    - Don't revalidate objects >= CHUNK_SIZE that aren't in the disk cache.
      Closes: #610199
  * drop patches included upstream and rebase the rest
  * update watch file

 -- Rolf Leggewie <email address hidden> Fri, 06 Jun 2014 15:29:19 +0200