Re: Planet stalling?

More info: for http://live.linuxchix.org/ (56 subscriptions, 8.2MB in
the cache directory according to du -sh) it uses up to 32MB of memory
while running. This does explain nicely why it's so painful to run on my
linode, which has a grand total of 64MB of RAM :(

-Mary

Re: Planet stalling?

Mary Gardiner wrote:

>More info: for http://live.linuxchix.org/ (56 subscriptions, 8.2MB in
>the cache directory according to du -sh) it uses up to 32MB of memory
>while running. This does explain nicely why it's so painful to run on my
>linode, which has a grand total of 64MB of RAM :(
>  
>
You may try to clear up the cache and let it get recreated from scratch.
It appears planetplanet has a bug with "old" caches, one of the 
manifestations is when you get a single blog filling up all of your Planet.

Simos

Re: Planet stalling?

On Thu, Sep 08, 2005, Mary Gardiner wrote:
> More info: for http://live.linuxchix.org/ (56 subscriptions, 8.2MB in
> the cache directory according to du -sh) it uses up to 32MB of memory
> while running. This does explain nicely why it's so painful to run on my
> linode, which has a grand total of 64MB of RAM :(

An off-list reply suggested that I remove my cache completely and re-run
Planet, because this effect is due to running Planet on an old cache.
Alas, this doesn't seem to be correct. On a completely empty cache, it
used about 29MB of memory.

-Mary

Re: Planet stalling?

>> More info: for http://live.linuxchix.org/ (56 subscriptions, 8.2MB in 
>> the cache directory according to du -sh) it uses up to 32MB of memory 
>> while running. This does explain nicely why it's so painful to run on 
>> my linode, which has a grand total of 64MB of RAM :(
>
> An off-list reply suggested that I remove my cache completely and re-run 
> Planet, because this effect is due to running Planet on an old cache. 
> Alas, this doesn't seem to be correct. On a completely empty cache, it 
> used about 29MB of memory.

... check to see if you might have a particular feed that's stalling or 
not fetching at a sane rate of speed.  using lsof to poke around for open 
sockets would be a place to start.

--elijah

Re: Planet stalling?

On Thu, Sep 08, 2005, elijah wright wrote:
> 
> >> More info: for http://live.linuxchix.org/ (56 subscriptions, 8.2MB in 
> >> the cache directory according to du -sh) it uses up to 32MB of memory 
> >> while running. This does explain nicely why it's so painful to run on 
> >> my linode, which has a grand total of 64MB of RAM :(
> >
> >An off-list reply suggested that I remove my cache completely and re-run 
> >Planet, because this effect is due to running Planet on an old cache. 
> >Alas, this doesn't seem to be correct. On a completely empty cache, it 
> >used about 29MB of memory.
> 
> ... check to see if you might have a particular feed that's stalling or 
> not fetching at a sane rate of speed.  using lsof to poke around for open 
> sockets would be a place to start.

Unless the system starts swapping heavily, which slows everything down
one hundred to one thousand fold, the process completes in three minutes
or so. What length of time should I be looking for something stalling
for?

The reason I'm concerned about this is that while in normal
circumstances, having it use half the available RAM for a few minutes is
OK, under any of the following circumstances it pushes my system into
swap hell:

 1. Apache is getting a lot of requests
 2. another intensive process (sa-learn) happens to run at the same time

This can easily cause enough problems with swapping (the Planet process

Re: Planet stalling?

Yesterday at 0:34, Mary Gardiner wrote:

> This can easily cause enough problems with swapping (the Planet process
> tends to just continually get swapped in and out in a dance with Apache)
> that the processes begin to take three to four hours to complete, which
> means that I've got three or four of them running at the same time,
> which means that it never resolves itself and I have to go in and start
> shutting processes down.

Then best would probably be to kill previous planet runs in a cron job
as well.  Earlier, when I used resource-limited machine as well (~64mb
RAM too), planetplanet caused other services to die occassionally. 

Moving to a more memory set-up improved everything, of course 

Cheers,
Danilo

【参加者すべてが儲かる仕組みです】

突然のメールで失礼致します。

事実です！参加者すべてが儲かる仕組みがあります。！
あなたも絶対に稼げるのです。

こんな情報に興味のある方は、http://flower.kakiko.com/mtn0806/index.html　をご覧ください。

突然のメール、大変失礼致しました。

link

Hello,

I have a Planet-powered website (Planet Wiki), how can I get a link to it from 
http://www.planetplanet.org/ ?

Thanks,
--

-- 
NSK
Founder of the Wikinerds Community
http://www.wikinerds.org/

Planet fix for content-type / atom 1.0

In Atom 1.0, the content type is set to
<content type="xhtml">

Trivial patch for planet to handle this is attached. Please apply.

-- 
Hanno Böck		Blog:   http://www.hboeck.de/
GPG: 3DBD3B20		Jabber: jabber <at> hboeck.de

In Atom 1.0, the content type is set to
<content type="xhtml">

Trivial patch for planet to handle this is attached. Please apply.

--

-- 
Hanno Böck		Blog:   http://www.hboeck.de/
GPG: 3DBD3B20		Jabber: jabber <at> hboeck.de

Re: Planet fix for content-type / atom 1.0

<quote who="Hanno Böck">

> In Atom 1.0, the content type is set to <content type="xhtml">
> 
> Trivial patch for planet to handle this is attached. Please apply.

Hi Hanno,

I'd love it if you could send this patch for feedparser upstream [1] - we
just include it in our tarball for ease of use.

Thanks,

- Jeff

[1] feedparser.sf.net appears to be the public location nowadays

--

-- 
EuroOSCON: October 17th-20th    http://conferences.oreillynet.com/eurooscon/

                       Wake up and smell the penguin.