| 1 |
You could try to limit processes with apache |
| 2 |
See apache manual for |
| 3 |
RLimitMEM and RLimitCPU. |
| 4 |
|
| 5 |
I'm not using Apache2 in production (or any) enviroment. Perhaps you could try |
| 6 |
apache-1.3.31 (latest one) and see if that gives you problems like apache2. |
| 7 |
|
| 8 |
regards, |
| 9 |
M. |
| 10 |
|
| 11 |
On Friday 14 May 2004 12:01, Raymond den Ouden wrote: |
| 12 |
> Oh sorry... |
| 13 |
> |
| 14 |
> I forgot to mention the server details |
| 15 |
> Apache: 2.0.49-r1 |
| 16 |
> Mod_PHP: 4.3.6-r1 |
| 17 |
> OpenSSL: 0.9.7c |
| 18 |
> Kernel: 2.4.26 |
| 19 |
> |
| 20 |
> Previously I had mod_mono and mailman enabled but I shutted that down this |
| 21 |
> week with no result. I also putted down mysql_log for apache, but for that |
| 22 |
> component I had to overwrite a two modules mod_log_config.so and |
| 23 |
> mod_logio.so . I am now compiling apache so that the modules are the |
| 24 |
> default ones. |
| 25 |
> |
| 26 |
> When looking around I ran into was DO NOT USE Apache 2 and PHP in a |
| 27 |
> production environment :( |
| 28 |
> Also I already tried to look at ulimit before you mentioned, but it did |
| 29 |
> not give really satisfying answers :( |
| 30 |
> |
| 31 |
> What could be a memory consuming item in the site might be: |
| 32 |
> - a dynamically generated image (1x1) pixel, so that should not be too |
| 33 |
> expensive for 800 pageviews a day, the script cleans up the mess itself. |
| 34 |
> - A download of 50 Meg, which was streamed by PHP for leechprotection at |
| 35 |
> the same time gzip was on, this could be a really memory consuming option. |
| 36 |
> - The CMS (Mambo Open Source 4.5 1.0.7) |
| 37 |
> |
| 38 |
> Well still trying to get it stable :( |
| 39 |
> |
| 40 |
> > Hi! |
| 41 |
> > |
| 42 |
> > On Fri, 14 May 2004, Raymond den Ouden wrote: |
| 43 |
> >> as I said a few days ago I launched a public web-site. after to days the |
| 44 |
> >> server was like a zombie. |
| 45 |
> > |
| 46 |
> > [... detective work ...] |
| 47 |
> > |
| 48 |
> >> But on killing apache2 the server came back and the command free told |
| 49 |
> >> me that 500 mb of swap and 350 MB was free-ed. |
| 50 |
> >> |
| 51 |
> >> I have looked around how I could strict apache in memory usage for the |
| 52 |
> >> apache user. |
| 53 |
> >> What is the best way to prevent that apache takes all memory or maybe |
| 54 |
> >> exclude a amount of memory for apache? |
| 55 |
> > |
| 56 |
> > Well, such memory hogging may have different causes. First off: |
| 57 |
> > about which exact version of Apache are we talking? Did you |
| 58 |
> > google about that version and possibly known memleaks? Other than |
| 59 |
> > that it's important to keep in mind that some forms of PHP and/or |
| 60 |
> > CGIs may have leaks that pile onto Apache's account. |
| 61 |
> > |
| 62 |
> > Beyond that, limiting memory usage may be done with ulimit. But I |
| 63 |
> > don't know how Apache handles (if at all) bumping into such a |
| 64 |
> > limit. |
| 65 |
> > |
| 66 |
> > I for one would rather try to find the source of the leak than |
| 67 |
> > just curing symptoms. Although I admit that getting your service |
| 68 |
> > up and running first might be feasible, remember that the |
| 69 |
> > underlying problem might be more grave and restarting your |
| 70 |
> > webservice might grind you into deeper trouble. |
| 71 |
> > |
| 72 |
> > Greets & HTH, |
| 73 |
> > Tobias |
| 74 |
> > -- |
| 75 |
> > Thank you for calling $PROVIDER helpdesk. If your cupholder is |
| 76 |
> > broken, please press 1. If you want an actual knowledgable support |
| 77 |
> > person, please enter the IP representation of a /28 netmask." |
| 78 |
> > |
| 79 |
> > -- |
| 80 |
> > gentoo-security@g.o mailing list |
| 81 |
> |
| 82 |
> -- |
| 83 |
> gentoo-security@g.o mailing list |
| 84 |
|
| 85 |
|
| 86 |
-- |
| 87 |
gentoo-security@g.o mailing list |
Archives