Urgent help needed: Zope falls over under moderate load
So yesterday my site got slashdotted. Zope fell over within minutes. A day later it's still falling over within an hour or so. I have 512Kb DSL (the outbound channel is 512Kb; the inbound channel is 768Kb). I figured that bandwidth would be saturated or that Windows 2000 Pro would melt long before Zope would fail, but that has proven not to be the case. Here's the particulars of my installation: Pentium III 850MHz with 768Mb RAM MS Windows 2000 Pro with current patches and service packs Zope 2.4.3 running as a service Squishdot 1.3.0 512Kb bandwidth Here's the registry start parameters: "C:\PROGRA~1\FARCES~1\bin\python.exe" "C:\PROGRA~1\FARCES~1\z2.py" -S -w80 -W9800 I don't know what the -S is; -w80 sets the port Zope's web server listens on; and -W9800 sets the WebDAV port. Zope just shuts down and closes port 80 on what seems to be very moderate load. We're talking a peak of less than 60,000 hits, a peak of 125Mb bandwidth usage, just over 58,000 page views, and a peak of 4013 visitors. What I've been doing is to manually stop the Zope service and restart it, but that's ridiculous. At this rate, clients are going to laugh at me when I suggest using Zope for a web project. Anyone have any ideas? I'm digestified, so please cc me in any response. Thanks kindly. -- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
Your clients maybe *should* laugh at you if you're not doing any caching. ;-) To cache, you have two options. Use Zope's RAM Cache Manager to do caching. As a result, the pages you choose to cache will be cached in RAM and won't be regenerated on each request. - OR - Use a caching web server in front of Zope to bear the brunt of the load in conjunction with (perhaps) the HTTP Accelerated Cache Manager stuff in Zope. In this case, your web server (or proxy) keeps a copy of a pregenerated page and serves it out to requestors instead of continually asking Zope to regenerate it. One of these is going to be necessary for any sort of high-traffic site. Zope can't keep up when it needs to dynamically generate content; you need to help it out by doing some kind of caching. If you don't or can't use Cache Managers, but you still want to use an HTTP caching proxy in front of your Zope site, you'll need to make each of the pages you want to cache has an "Expires" and "Last Modified" header. You may do this by using RESPONSE.setHeader if you're not using a cache manager. If you're behind Apache, you can use mod_proxy with caching turned on. See http://www.zope.org/SiteIndex/search?text_content=caching for more information. I've also found the document about mod_perl and headers at http://perl.apache.org/guide/correct_headers.html helpful... there are lots of other nuggets in that document as well. HTH, - C ----- Original Message ----- From: "Michael Fraase" <mfraase@farces.com> To: <zope@zope.org> Sent: Tuesday, November 20, 2001 12:39 PM Subject: [Zope] Urgent help needed: Zope falls over under moderate load
So yesterday my site got slashdotted. Zope fell over within minutes. A day later it's still falling over within an hour or so.
I have 512Kb DSL (the outbound channel is 512Kb; the inbound channel is 768Kb). I figured that bandwidth would be saturated or that Windows 2000 Pro would melt long before Zope would fail, but that has proven not to be the case.
Here's the particulars of my installation:
Pentium III 850MHz with 768Mb RAM MS Windows 2000 Pro with current patches and service packs Zope 2.4.3 running as a service Squishdot 1.3.0 512Kb bandwidth
Here's the registry start parameters:
"C:\PROGRA~1\FARCES~1\bin\python.exe" "C:\PROGRA~1\FARCES~1\z2.py" -S -w80 -W9800
I don't know what the -S is; -w80 sets the port Zope's web server listens on; and -W9800 sets the WebDAV port.
Zope just shuts down and closes port 80 on what seems to be very moderate load. We're talking a peak of less than 60,000 hits, a peak of 125Mb bandwidth usage, just over 58,000 page views, and a peak of 4013 visitors.
What I've been doing is to manually stop the Zope service and restart it, but that's ridiculous. At this rate, clients are going to laugh at me when I suggest using Zope for a web project.
Anyone have any ideas? I'm digestified, so please cc me in any response.
Thanks kindly.
-- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )
Thanks Chris, I've cached what I can find. The only page that's cacheable is index_html; the pages being hit are another layer deeper in the hierarchy. I've also cached all of the page elements (images, header, footer) I could find. -- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
-----Original Message----- From: Chris McDonough [mailto:chrism@zope.com] Sent: Tuesday, November 20, 2001 12:11 PM To: mfraase@farces.com; zope@zope.org Subject: Re: [Zope] Urgent help needed: Zope falls over under moderate load
Your clients maybe *should* laugh at you if you're not doing any caching. ;-)
To cache, you have two options.
Use Zope's RAM Cache Manager to do caching. As a result, the pages you choose to cache will be cached in RAM and won't be regenerated on each request.
- OR -
Use a caching web server in front of Zope to bear the brunt of the load in conjunction with (perhaps) the HTTP Accelerated Cache Manager stuff in Zope. In this case, your web server (or proxy) keeps a copy of a pregenerated page and serves it out to requestors instead of continually asking Zope to regenerate it.
One of these is going to be necessary for any sort of high-traffic site. Zope can't keep up when it needs to dynamically generate content; you need to help it out by doing some kind of caching.
If you don't or can't use Cache Managers, but you still want to use an HTTP caching proxy in front of your Zope site, you'll need to make each of the pages you want to cache has an "Expires" and "Last Modified" header. You may do this by using RESPONSE.setHeader if you're not using a cache manager. If you're behind Apache, you can use mod_proxy with caching turned on.
See http://www.zope.org/SiteIndex/search?text_content=caching for more information.
I've also found the document about mod_perl and headers at http://perl.apache.org/guide/correct_headers.html helpful... there are lots of other nuggets in that document as well.
HTH,
- C
----- Original Message ----- From: "Michael Fraase" <mfraase@farces.com> To: <zope@zope.org> Sent: Tuesday, November 20, 2001 12:39 PM Subject: [Zope] Urgent help needed: Zope falls over under moderate load
So yesterday my site got slashdotted. Zope fell over within minutes. A day later it's still falling over within an hour or so.
I have 512Kb DSL (the outbound channel is 512Kb; the inbound channel is 768Kb). I figured that bandwidth would be saturated or that Windows 2000 Pro would melt long before Zope would fail, but that has proven not to be the case.
Here's the particulars of my installation:
Pentium III 850MHz with 768Mb RAM MS Windows 2000 Pro with current patches and service packs Zope 2.4.3 running as a service Squishdot 1.3.0 512Kb bandwidth
Here's the registry start parameters:
"C:\PROGRA~1\FARCES~1\bin\python.exe" "C:\PROGRA~1\FARCES~1\z2.py" -S -w80 -W9800
I don't know what the -S is; -w80 sets the port Zope's web server listens on; and -W9800 sets the WebDAV port.
Zope just shuts down and closes port 80 on what seems to be very moderate load. We're talking a peak of less than 60,000 hits, a peak of 125Mb bandwidth usage, just over 58,000 page views, and a peak of 4013 visitors.
What I've been doing is to manually stop the Zope service and restart it, but that's ridiculous. At this rate, clients are going to laugh at me when I suggest using Zope for a web project.
Anyone have any ideas? I'm digestified, so please cc me in any response.
Thanks kindly.
-- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce
Are the pages uncacheable because they're restricted by security or for another reason? ----- Original Message ----- From: "Michael Fraase" <mfraase@farces.com> To: "'Chris McDonough'" <chrism@zope.com>; <zope@zope.org> Sent: Tuesday, November 20, 2001 1:14 PM Subject: RE: [Zope] Urgent help needed: Zope falls over under moderate load
Thanks Chris,
I've cached what I can find. The only page that's cacheable is index_html; the pages being hit are another layer deeper in the hierarchy. I've also cached all of the page elements (images, header, footer) I could find.
-- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
-----Original Message----- From: Chris McDonough [mailto:chrism@zope.com] Sent: Tuesday, November 20, 2001 12:11 PM To: mfraase@farces.com; zope@zope.org Subject: Re: [Zope] Urgent help needed: Zope falls over under moderate load
Your clients maybe *should* laugh at you if you're not doing any caching. ;-)
To cache, you have two options.
Use Zope's RAM Cache Manager to do caching. As a result, the pages you choose to cache will be cached in RAM and won't be regenerated on each request.
- OR -
Use a caching web server in front of Zope to bear the brunt of the load in conjunction with (perhaps) the HTTP Accelerated Cache Manager stuff in Zope. In this case, your web server (or proxy) keeps a copy of a pregenerated page and serves it out to requestors instead of continually asking Zope to regenerate it.
One of these is going to be necessary for any sort of high-traffic site. Zope can't keep up when it needs to dynamically generate content; you need to help it out by doing some kind of caching.
If you don't or can't use Cache Managers, but you still want to use an HTTP caching proxy in front of your Zope site, you'll need to make each of the pages you want to cache has an "Expires" and "Last Modified" header. You may do this by using RESPONSE.setHeader if you're not using a cache manager. If you're behind Apache, you can use mod_proxy with caching turned on.
See http://www.zope.org/SiteIndex/search?text_content=caching for more information.
I've also found the document about mod_perl and headers at http://perl.apache.org/guide/correct_headers.html helpful... there are lots of other nuggets in that document as well.
HTH,
- C
----- Original Message ----- From: "Michael Fraase" <mfraase@farces.com> To: <zope@zope.org> Sent: Tuesday, November 20, 2001 12:39 PM Subject: [Zope] Urgent help needed: Zope falls over under moderate load
So yesterday my site got slashdotted. Zope fell over within minutes. A day later it's still falling over within an hour or so.
I have 512Kb DSL (the outbound channel is 512Kb; the inbound channel is 768Kb). I figured that bandwidth would be saturated or that Windows 2000 Pro would melt long before Zope would fail, but that has proven not to be the case.
Here's the particulars of my installation:
Pentium III 850MHz with 768Mb RAM MS Windows 2000 Pro with current patches and service packs Zope 2.4.3 running as a service Squishdot 1.3.0 512Kb bandwidth
Here's the registry start parameters:
"C:\PROGRA~1\FARCES~1\bin\python.exe" "C:\PROGRA~1\FARCES~1\z2.py" -S -w80 -W9800
I don't know what the -S is; -w80 sets the port Zope's web server listens on; and -W9800 sets the WebDAV port.
Zope just shuts down and closes port 80 on what seems to be very moderate load. We're talking a peak of less than 60,000 hits, a peak of 125Mb bandwidth usage, just over 58,000 page views, and a peak of 4013 visitors.
What I've been doing is to manually stop the Zope service and restart it, but that's ridiculous. At this rate, clients are going to laugh at me when I suggest using Zope for a web project.
Anyone have any ideas? I'm digestified, so please cc me in any response.
Thanks kindly.
-- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )
They just don't show up in the list of cacheable items. -- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
-----Original Message----- From: Chris McDonough [mailto:chrism@zope.com] Sent: Tuesday, November 20, 2001 12:45 PM To: mfraase@farces.com; zope@zope.org Subject: Re: [Zope] Urgent help needed: Zope falls over under moderate load
Are the pages uncacheable because they're restricted by security or for another reason?
----- Original Message ----- From: "Michael Fraase" <mfraase@farces.com> To: "'Chris McDonough'" <chrism@zope.com>; <zope@zope.org> Sent: Tuesday, November 20, 2001 1:14 PM Subject: RE: [Zope] Urgent help needed: Zope falls over under moderate load
Thanks Chris,
I've cached what I can find. The only page that's cacheable is index_html; the pages being hit are another layer deeper in the hierarchy. I've also cached all of the page elements (images, header, footer) I could find.
-- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
-----Original Message----- From: Chris McDonough [mailto:chrism@zope.com] Sent: Tuesday, November 20, 2001 12:11 PM To: mfraase@farces.com; zope@zope.org Subject: Re: [Zope] Urgent help needed: Zope falls over under moderate load
Your clients maybe *should* laugh at you if you're not doing any caching. ;-)
To cache, you have two options.
Use Zope's RAM Cache Manager to do caching. As a result, the pages you choose to cache will be cached in RAM and won't be regenerated on each request.
- OR -
Use a caching web server in front of Zope to bear the brunt of the load in conjunction with (perhaps) the HTTP Accelerated Cache Manager stuff in Zope. In this case, your web server (or proxy) keeps a copy of a pregenerated page and serves it out to requestors instead of continually asking Zope to regenerate it.
One of these is going to be necessary for any sort of high-traffic site. Zope can't keep up when it needs to dynamically generate content; you need to help it out by doing some kind of caching.
If you don't or can't use Cache Managers, but you still want to use an HTTP caching proxy in front of your Zope site, you'll need to make each of the pages you want to cache has an "Expires" and "Last Modified" header. You may do this by using RESPONSE.setHeader if you're not using a cache manager. If you're behind Apache, you can use mod_proxy with caching turned on.
See http://www.zope.org/SiteIndex/search?text_content=caching for more information.
I've also found the document about mod_perl and headers at http://perl.apache.org/guide/correct_headers.html helpful... there are lots of other nuggets in that document as well.
HTH,
- C
----- Original Message ----- From: "Michael Fraase" <mfraase@farces.com> To: <zope@zope.org> Sent: Tuesday, November 20, 2001 12:39 PM Subject: [Zope] Urgent help needed: Zope falls over under moderate load
So yesterday my site got slashdotted. Zope fell over within minutes. A day later it's still falling over within an hour or so.
I have 512Kb DSL (the outbound channel is 512Kb; the inbound channel is 768Kb). I figured that bandwidth would be saturated or that Windows 2000 Pro would melt long before Zope would fail, but that has proven not to be the case.
Here's the particulars of my installation:
Pentium III 850MHz with 768Mb RAM MS Windows 2000 Pro with current patches and service packs Zope 2.4.3 running as a service Squishdot 1.3.0 512Kb bandwidth
Here's the registry start parameters:
"C:\PROGRA~1\FARCES~1\bin\python.exe" "C:\PROGRA~1\FARCES~1\z2.py" -S -w80 -W9800
I don't know what the -S is; -w80 sets the port Zope's web server listens on; and -W9800 sets the WebDAV port.
Zope just shuts down and closes port 80 on what seems to be very moderate load. We're talking a peak of less than 60,000 hits, a peak of 125Mb bandwidth usage, just over 58,000 page views, and a peak of 4013 visitors.
What I've been doing is to manually stop the Zope service and restart it, but that's ridiculous. At this rate, clients are going to laugh at me when I suggest using Zope for a web project.
Anyone have any ideas? I'm digestified, so please cc me in any response.
Thanks kindly.
-- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce
This might be due to a configuration incompatibility with the Cache Manager stuff (for example, ZClass methods can't be cached). As a result, you may need to cache the pages manually by using a reverse caching proxy and Expires/Last-Modified headers as I described. I realize that needing to go in and tweak your configuration is not something you want to have to do while your site is getting hammered, but there's probably not much of a choice. That said, it's likely that the emergency will be over with before you get a working configuration up, which is of course frustrating. Hopefully, at some point we'll have a pretty and comprehensive set of step-by-step documents describing specifically how to set Zope up for high load. In the meantime, the rule of thumb is "always cache as much as possible"; the corollary is "always test your cache hit rate before putting something into production". You'll probably want to do both before you get Slashdotted the next time. ----- Original Message ----- From: "Michael Fraase" <mfraase@farces.com> To: "'Chris McDonough'" <chrism@zope.com>; <zope@zope.org> Sent: Tuesday, November 20, 2001 2:22 PM Subject: RE: [Zope] Urgent help needed: Zope falls over under moderate load
They just don't show up in the list of cacheable items.
-- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
-----Original Message----- From: Chris McDonough [mailto:chrism@zope.com] Sent: Tuesday, November 20, 2001 12:45 PM To: mfraase@farces.com; zope@zope.org Subject: Re: [Zope] Urgent help needed: Zope falls over under moderate load
Are the pages uncacheable because they're restricted by security or for another reason?
----- Original Message ----- From: "Michael Fraase" <mfraase@farces.com> To: "'Chris McDonough'" <chrism@zope.com>; <zope@zope.org> Sent: Tuesday, November 20, 2001 1:14 PM Subject: RE: [Zope] Urgent help needed: Zope falls over under moderate load
Thanks Chris,
I've cached what I can find. The only page that's cacheable is index_html; the pages being hit are another layer deeper in the hierarchy. I've also cached all of the page elements (images, header, footer) I could find.
-- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
-----Original Message----- From: Chris McDonough [mailto:chrism@zope.com] Sent: Tuesday, November 20, 2001 12:11 PM To: mfraase@farces.com; zope@zope.org Subject: Re: [Zope] Urgent help needed: Zope falls over under moderate load
Your clients maybe *should* laugh at you if you're not doing any caching. ;-)
To cache, you have two options.
Use Zope's RAM Cache Manager to do caching. As a result, the pages you choose to cache will be cached in RAM and won't be regenerated on each request.
- OR -
Use a caching web server in front of Zope to bear the brunt of the load in conjunction with (perhaps) the HTTP Accelerated Cache Manager stuff in Zope. In this case, your web server (or proxy) keeps a copy of a pregenerated page and serves it out to requestors instead of continually asking Zope to regenerate it.
One of these is going to be necessary for any sort of high-traffic site. Zope can't keep up when it needs to dynamically generate content; you need to help it out by doing some kind of caching.
If you don't or can't use Cache Managers, but you still want to use an HTTP caching proxy in front of your Zope site, you'll need to make each of the pages you want to cache has an "Expires" and "Last Modified" header. You may do this by using RESPONSE.setHeader if you're not using a cache manager. If you're behind Apache, you can use mod_proxy with caching turned on.
See http://www.zope.org/SiteIndex/search?text_content=caching for more information.
I've also found the document about mod_perl and headers at http://perl.apache.org/guide/correct_headers.html helpful... there are lots of other nuggets in that document as well.
HTH,
- C
----- Original Message ----- From: "Michael Fraase" <mfraase@farces.com> To: <zope@zope.org> Sent: Tuesday, November 20, 2001 12:39 PM Subject: [Zope] Urgent help needed: Zope falls over under moderate load
So yesterday my site got slashdotted. Zope fell over within minutes. A day later it's still falling over within an hour or so.
I have 512Kb DSL (the outbound channel is 512Kb; the inbound channel is 768Kb). I figured that bandwidth would be saturated or that Windows 2000 Pro would melt long before Zope would fail, but that has proven not to be the case.
Here's the particulars of my installation:
Pentium III 850MHz with 768Mb RAM MS Windows 2000 Pro with current patches and service packs Zope 2.4.3 running as a service Squishdot 1.3.0 512Kb bandwidth
Here's the registry start parameters:
"C:\PROGRA~1\FARCES~1\bin\python.exe" "C:\PROGRA~1\FARCES~1\z2.py" -S -w80 -W9800
I don't know what the -S is; -w80 sets the port Zope's web server listens on; and -W9800 sets the WebDAV port.
Zope just shuts down and closes port 80 on what seems to be very moderate load. We're talking a peak of less than 60,000 hits, a peak of 125Mb bandwidth usage, just over 58,000 page views, and a peak of 4013 visitors.
What I've been doing is to manually stop the Zope service and restart it, but that's ridiculous. At this rate, clients are going to laugh at me when I suggest using Zope for a web project.
Anyone have any ideas? I'm digestified, so please cc me in any response.
Thanks kindly.
-- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )
Hi, Am Tue, 20 Nov 2001 12:14:23 -0600 schrieb Michael Fraase:
I've cached what I can find. The only page that's cacheable is index_html; the pages being hit are another layer deeper in the hierarchy. I've also cached all of the page elements (images, header, footer) I could find.
When creating a own page class (using a product), what should I do that the cache managers can cache those pages? Gruß, Markus -- You don't have to be Microsoft to suck... but it helps. (Tim Hammerquist in comp.lang.python)
Chris McDonough wrote:
Your clients maybe *should* laugh at you if you're not doing any caching. ;-)
Hurm... so Zope's no good if you want to build a scalable _and_ dynamic site? surely that's a little worrying? Chris
Hurm... so Zope's no good if you want to build a scalable _and_ dynamic site?
surely that's a little worrying?
Not really. Zope depends on caching and cannot survive a high transaction rate without it unless your site is very very very simple. Usually, if you don't cache or serve only static content, you can't running a high-volume site using Zope or mod_perl or ASP or whatever-else; it's as simple as that. But you knew that. - C
I guess I'm confused. Everything that *could* be cached *was* cached. And no, I don't run a caching server or a proxy server or anything else in front of Zope. I'm a writer, not a programmer. The /. piece hit about 1:00 AM. By 1:01 AM Zope had folded like a cheap suit. It's still going down about every 40 minutes or so. Now remember, my outbound bandwidth is limited to 512Kb. Am I correct in my understanding that Zope can't handle even 512Kb of demand without some technical doohickey in front of it so it doesn't fall down? No offense intended, but I think two internal Squishdot pages meet the definition of pretty dang simple. And why does it fall over anyway? This just doesn't make any sense to me. I can see it getting slow and timing out, but giving up completely and just bailing? What's that about? Explain it to me like I'm an intelligent, non-technical friend. Thanks. -- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
-----Original Message----- From: Chris McDonough [mailto:chrism@zope.com] Sent: Tuesday, November 20, 2001 2:38 PM To: Chris Withers Cc: mfraase@farces.com; zope@zope.org Subject: Re: [Zope] Urgent help needed: Zope falls over under moderate load
Hurm... so Zope's no good if you want to build a scalable _and_ dynamic site?
surely that's a little worrying?
Not really. Zope depends on caching and cannot survive a high transaction rate without it unless your site is very very very simple. Usually, if you don't cache or serve only static content, you can't running a high-volume site using Zope or mod_perl or ASP or whatever-else; it's as simple as that.
But you knew that.
- C
Chris McDonough wrote:
Usually, if you don't cache or serve only static content, you can't running a high-volume site using Zope or mod_perl or ASP or whatever-else; it's as simple as that.
What about something like Google? How do they get around it? I'm sure caching isn't the only option... Chris
What about something like Google? How do they get around it?
I'd be surprised if they didn't use some form of caching. Do you know that they don't?
I'm sure caching isn't the only option...
It's not the only option, but it's the fastest, surest, and most cost-effective one.
At 21:06 20-11-2001 +0000, Chris Withers wrote:
Chris McDonough wrote:
Usually, if you don't cache or serve only static content, you can't running a high-volume site using Zope or mod_perl or ASP or whatever-else; it's as simple as that.
What about something like Google? How do they get around it?
I'm sure caching isn't the only option...
This discussion is so pointless... FYI Google has upwards of 10.000 (thats TEN THOUSAND) linux machines serving as frontends for their service. And although Google's pages are dynamic, they're basically the same script run over and over again with little or no images. So caching is not the only option; its just that I'm sure you dont have Google's budget. What you're trying to do is asking Zope to DYNAMICALLY generate Squishdot pages (probably full of database accesses and images). But you're doing this on a PC (no hardware specs were specified) running Windows (!) and on a DSL line. If you try to do something similar with ANY OTHER platform (lets say Apache+PHP+PHPNuke; or IIS+ASP+some forum software or other) you're also going to get into the same kind of problems. If you compound this with the fact that you're also clueless (e.g running Zope without knowing what -S is for; or, for that matter, running any software without knowing what you're doing), you're in for a lot of trouble. Please, dont answer "I'm a writer, not a programmer". THAT is precisely your problem. You want to be able to build rockets without taking a course in rocket science. Dont misunderstand me; you can still fire rockets or build amateur rockets; just dont try to reach the moon. Here are some tips: - enable caching on Zope - change from Windows to Linux - install more RAM on your PC - get more CPU for your PC - put Squid in front of your Zope - fine tune your Zope to use only the necessary command line switches - pack your Zope database frequently - stop all the other software that you might have running (i.e. dedicated server) - create static versions of the most accessed page(s) - you might have a memory leak somewhere; plug it We're running a consumer portal with about 1 million hits per day using 2 frontends (regular PC machines) connected to 1 backend (server PC). C U! -- Mario Valente
Perhaps he is having a crisis in his personal life and feels belittling people will help his own ego problems. Or maybe he's just wantonly justified in everything. Perhaps we should address him as 'master'... Paul Zwarts -----Original Message----- From: zope-admin@zope.org [mailto:zope-admin@zope.org] On Behalf Of Chris Withers Sent: Thursday, November 22, 2001 1:06 PM To: Mario Valente Cc: zope@zope.org Subject: [Zope] :-S Mario Valente wrote:
This discussion is so pointless...
I have to say, I think you're reply was pretty rude and unhelpful, even by my standards ;-) Chris _______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )
At 13:23 22-11-2001 +0100, Paul Zwarts wrote:
Perhaps he is having a crisis in his personal life and feels belittling people will help his own ego problems. Or maybe he's just wantonly justified in everything. Perhaps we should address him as 'master'...
Actually, "Generalissimo" will do... And I'm not "wantonly justified in everything". Just when people are making uninformed aggressive questions and using uninformed and baseless facts to keep on being uninformed just as long as they dont have to do their homework and someone else solves their problem out of their interest in being kind. C U! -- Mario Valente
Ah.. I can concur with that... the absence of mind is frustrating to self-proclaimed brilliance. I am one of those too... ; ) Va bene, Paul Zwarts -----Original Message----- From: zope-admin@zope.org [mailto:zope-admin@zope.org] On Behalf Of Mario Valente Sent: Thursday, November 22, 2001 1:20 PM To: zope@zope.org Subject: RE: [Zope] :-S At 13:23 22-11-2001 +0100, Paul Zwarts wrote:
Perhaps he is having a crisis in his personal life and feels belittling people will help his own ego problems. Or maybe he's just wantonly justified in everything. Perhaps we should address him as 'master'...
Actually, "Generalissimo" will do... And I'm not "wantonly justified in everything". Just when people are making uninformed aggressive questions and using uninformed and baseless facts to keep on being uninformed just as long as they dont have to do their homework and someone else solves their problem out of their interest in being kind. C U! -- Mario Valente _______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )
Come on, folks, chill out, if you please! Tom P
Chris Withers writes:
Dieter Maurer wrote:
Paul Zwarts writes:
Ah.. I can concur with that... the absence of mind is frustrating to self-proclaimed brilliance. I am one of those too... ; ) Can we stop this thread, please....
No ;-) Sad that we have time to flame one another...
Dieter
Aw, come on... Didn't you know that Turkey contains a chemical ingredient that causes one to feel slow, stupid and lazy? Its all fun and games until someone looses an eye. Then it's a sport! 8:-) Paz -----Original Message----- From: zope-admin@zope.org [mailto:zope-admin@zope.org] On Behalf Of Dieter Maurer Sent: Friday, November 23, 2001 6:04 PM To: Chris Withers Cc: Paul Zwarts; 'Mario Valente'; zope@zope.org Subject: Re: [Zope] :-S Chris Withers writes:
Dieter Maurer wrote:
Paul Zwarts writes:
Ah.. I can concur with that... the absence of mind is
frustrating to
self-proclaimed brilliance. I am one of those too... ; ) Can we stop this thread, please....
No ;-) Sad that we have time to flame one another...
Dieter _______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )
What about the -t option, to create more threads?
From: "Michael Fraase" <mfraase@farces.com> Organization: ARTS & FARCES LLC Reply-To: <mfraase@farces.com> Date: Tue, 20 Nov 2001 14:40:25 -0600 To: "'Chris McDonough'" <chrism@zope.com>, "'Chris Withers'" <chrisw@nipltd.com> Cc: <zope@zope.org> Subject: RE: [Zope] Urgent help needed: Zope falls over under moderate load
I guess I'm confused. Everything that *could* be cached *was* cached. And no, I don't run a caching server or a proxy server or anything else in front of Zope. I'm a writer, not a programmer.
The /. piece hit about 1:00 AM. By 1:01 AM Zope had folded like a cheap suit. It's still going down about every 40 minutes or so.
Now remember, my outbound bandwidth is limited to 512Kb.
Am I correct in my understanding that Zope can't handle even 512Kb of demand without some technical doohickey in front of it so it doesn't fall down?
No offense intended, but I think two internal Squishdot pages meet the definition of pretty dang simple.
And why does it fall over anyway? This just doesn't make any sense to me. I can see it getting slow and timing out, but giving up completely and just bailing? What's that about? Explain it to me like I'm an intelligent, non-technical friend. Thanks.
-- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
-----Original Message----- From: Chris McDonough [mailto:chrism@zope.com] Sent: Tuesday, November 20, 2001 2:38 PM To: Chris Withers Cc: mfraase@farces.com; zope@zope.org Subject: Re: [Zope] Urgent help needed: Zope falls over under moderate load
Hurm... so Zope's no good if you want to build a scalable _and_ dynamic site?
surely that's a little worrying?
Not really. Zope depends on caching and cannot survive a high transaction rate without it unless your site is very very very simple. Usually, if you don't cache or serve only static content, you can't running a high-volume site using Zope or mod_perl or ASP or whatever-else; it's as simple as that.
But you knew that.
- C
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )
Michael Fraase wrote:
Am I correct in my understanding that Zope can't handle even 512Kb of demand without some technical doohickey in front of it so it doesn't fall down?
Well, maybe. I agree it shouldn't die. It's a shame you run it as a service, I have no idea where the output from those gets piped to. If you run it in a DOS box, you can see Zope starting up, and dying ;-) It's usually polite enough to fart just before it dies, which can give a lot of clues...
No offense intended, but I think two internal Squishdot pages meet the definition of pretty dang simple.
I'll take that as a compliment, thankyou :-) Two internal Squishdot pages are not neccesarily that dang simple to deliver, it just appears that way 'cos the app is nice and easy to set up...
And why does it fall over anyway?
Start sniffing zope farts and then maybe you can find out :-) hoping-to-help, Chris :-)
On Wednesday 21 November 2001 08:10, Chris Withers wrote:
Michael Fraase wrote:
Am I correct in my understanding that Zope can't handle even 512Kb of demand without some technical doohickey in front of it so it doesn't fall down?
Well, maybe. I agree it shouldn't die. It's a shame you run it as a service, I have no idea where the output from those gets piped to. If you run it in a DOS box, you can see Zope starting up, and dying ;-) It's usually polite enough to fart just before it dies, which can give a lot of clues...
You could also try launching from the Zope Controller - it logs all output. http://www.zope.org/Members/richard/zope_controller/win It also uses a different, lighter-weight NT service. Richard
hello,
Am I correct in my understanding that Zope can't handle even 512Kb of demand without some technical doohickey in front of it so it doesn't fall down?
your problem may be w2k pro, not zope. w2k pro is limited to 5 simultaneous connections. your problem may also be your dsl line. i've had issues serving from dsl. i've hammered a zope box where it was serving several million pages a day, but very slowly. pages were still getting through though. mind you, that was not on a dsl line, but on a lan. luke
I guess I'm confused. Everything that *could* be cached *was* cached. And no, I don't run a caching server or a proxy server or anything else in front of Zope. I'm a writer, not a programmer.
OK, fair enough. But your profession still doesn't absolve you from needing to cache more in order to survive a Slashdotting. ;-) Either that or you'll need to start developing your site with static pages only. That'd work too.
The /. piece hit about 1:00 AM. By 1:01 AM Zope had folded like a cheap suit. It's still going down about every 40 minutes or so.
Now remember, my outbound bandwidth is limited to 512Kb.
If 512Kb/s is hit by as many 300-byte requests per minute as possible, this translates into without taking into account latency or response usage a potential inbound rate of 213 requests per second. That's still a lot of requests. As something to measure that up against at peak normal load, Slashdot gets about 180 requests/sec. The 512Kb/s isn't much of a throttle. And this is assuming that your inbound bandwidth is limited to 512Kb/s.. you only mentioned your outbound in this mail. If inbound is higher, it's even more of a problem.
Am I correct in my understanding that Zope can't handle even 512Kb of demand without some technical doohickey in front of it so it doesn't fall down?
Your pipe is fat enough to allow lots of requests in, and what you're serving is probably sufficiently complex to be very slow. Squishdot is really not known for its speed. "Raw" Zope itself could almost certainly handle it, however, if what you were returning is a DTML method that said "<html>this is a simple page</html>". But this isn't what you're returning; Squishdot has a big say in what shows up.
No offense intended, but I think two internal Squishdot pages meet the definition of pretty dang simple.
Maybe conceptually it's simple, but apps like Squishdot do lots of stuff in order to generate these pages. For fun, you should try to set up a "barebones" Squishot with the default homepage, and hit it repeatedly with a load-generator like Apache's "ab". Then try the same thing with a Zope page that is "<html>Hello!</html>". You will see a big difference. On an 850Mhz box at ZC, I can get Zope to serve about 152 requests/s with the simple page. Anybody want to try this with an out of the box Squishdot homepage? Or a Squishdot story page? The guy from the KDE dot (http://dot.kde.org) claimed he could only get about 2 requests/second out of a Squishdot home page. After setting up caching properly, he was able to get about 2000.
And why does it fall over anyway? This just doesn't make any sense to me. I can see it getting slow and timing out, but giving up completely and just bailing? What's that about? Explain it to me like I'm an intelligent, non-technical friend. Thanks.
The big "bang for buck" solution provider is caching. Assuming that you had no problems *before* the slashdotting, that will solve your problem because it will cause Zope to need to serve far fewer requests, closer to the number of requests you normally get. And this is (I assume) the outcome that you actually want. I highly recommend setting up a caching proxy in front of Zope if this sort of load will be recurring. It's way faster and cheaper than trying to understand the problem deeply. ;-) Most commercial sites are developed using this principle, AFAICT. But if you're as interested in understanding the phenomena as you are in solving the problem and you'd like to help the current Squishdot maintainer and ZC improve their products' behavior under load, it'd be necessary to know more details about how it was failing under load and what happened during the failures. I would be interested in these results. It could be a memory leak, it could be a Zope bug, a Squishdot bug, it could be just about anything. You need forensic information and you need to let it fail under load in order to get it. Usually, you can get this info by turning on "big M" logging (by passing "-M detailed.log" at the end of your start.bat script, maybe). On Linux, I'd recommend also using the ForensicLogger product (see http://www.zope.org/Members/mcdonc) to gather more details such as memory utilization and CPU utilization; it doesn't work on Windows, however. If you're willing to do this, let it fail under load, then send the log with the failure in it to me and I will try to analyze it. Note that you *might* be able to make use of the AutoLance product at http://www.zope.org/Members/mcdonc to autorestart your machine for you if you've got a memory leak. HTH, - C
Okay, Chris, that's a lot to chew on. Thanks very much for your patience and time. -- Michael Fraase ARTS & FARCES LLC mfraase@farces.com www.farces.com PGP Fingerprint: 3D85 F3F4 9E65 4949 176A 260C CB47 190D C864 9A96
-----Original Message----- From: Chris McDonough [mailto:chrism@zope.com] Sent: Tuesday, November 20, 2001 3:39 PM To: mfraase@farces.com; 'Chris Withers' Cc: zope@zope.org Subject: Re: [Zope] Urgent help needed: Zope falls over under moderate load
I guess I'm confused. Everything that *could* be cached *was* cached. And no, I don't run a caching server or a proxy server or anything else in front of Zope. I'm a writer, not a programmer.
OK, fair enough.
But your profession still doesn't absolve you from needing to cache more in order to survive a Slashdotting. ;-) Either that or you'll need to start developing your site with static pages only. That'd work too.
The /. piece hit about 1:00 AM. By 1:01 AM Zope had folded like a cheap suit. It's still going down about every 40 minutes or so.
Now remember, my outbound bandwidth is limited to 512Kb.
If 512Kb/s is hit by as many 300-byte requests per minute as possible, this translates into without taking into account latency or response usage a potential inbound rate of 213 requests per second. That's still a lot of requests. As something to measure that up against at peak normal load, Slashdot gets about 180 requests/sec. The 512Kb/s isn't much of a throttle.
And this is assuming that your inbound bandwidth is limited to 512Kb/s.. you only mentioned your outbound in this mail. If inbound is higher, it's even more of a problem.
Am I correct in my understanding that Zope can't handle even 512Kb of demand without some technical doohickey in front of it so it doesn't fall down?
Your pipe is fat enough to allow lots of requests in, and what you're serving is probably sufficiently complex to be very slow. Squishdot is really not known for its speed.
"Raw" Zope itself could almost certainly handle it, however, if what you were returning is a DTML method that said "<html>this is a simple page</html>". But this isn't what you're returning; Squishdot has a big say in what shows up.
No offense intended, but I think two internal Squishdot pages meet the definition of pretty dang simple.
Maybe conceptually it's simple, but apps like Squishdot do lots of stuff in order to generate these pages. For fun, you should try to set up a "barebones" Squishot with the default homepage, and hit it repeatedly with a load-generator like Apache's "ab". Then try the same thing with a Zope page that is "<html>Hello!</html>". You will see a big difference. On an 850Mhz box at ZC, I can get Zope to serve about 152 requests/s with the simple page.
Anybody want to try this with an out of the box Squishdot homepage? Or a Squishdot story page? The guy from the KDE dot (http://dot.kde.org) claimed he could only get about 2 requests/second out of a Squishdot home page. After setting up caching properly, he was able to get about 2000.
And why does it fall over anyway? This just doesn't make any sense to me. I can see it getting slow and timing out, but giving up completely and just bailing? What's that about? Explain it to me like I'm an intelligent, non-technical friend. Thanks.
The big "bang for buck" solution provider is caching. Assuming that you had no problems *before* the slashdotting, that will solve your problem because it will cause Zope to need to serve far fewer requests, closer to the number of requests you normally get. And this is (I assume) the outcome that you actually want. I highly recommend setting up a caching proxy in front of Zope if this sort of load will be recurring. It's way faster and cheaper than trying to understand the problem deeply. ;-) Most commercial sites are developed using this principle, AFAICT.
But if you're as interested in understanding the phenomena as you are in solving the problem and you'd like to help the current Squishdot maintainer and ZC improve their products' behavior under load, it'd be necessary to know more details about how it was failing under load and what happened during the failures. I would be interested in these results. It could be a memory leak, it could be a Zope bug, a Squishdot bug, it could be just about anything. You need forensic information and you need to let it fail under load in order to get it.
Usually, you can get this info by turning on "big M" logging (by passing "-M detailed.log" at the end of your start.bat script, maybe). On Linux, I'd recommend also using the ForensicLogger product (see http://www.zope.org/Members/mcdonc) to gather more details such as memory utilization and CPU utilization; it doesn't work on Windows, however. If you're willing to do this, let it fail under load, then send the log with the failure in it to me and I will try to analyze it.
Note that you *might* be able to make use of the AutoLance product at http://www.zope.org/Members/mcdonc to autorestart your machine for you if you've got a memory leak.
HTH,
- C
On Wednesday 21 November 2001 05:38 am, you wrote:
The big "bang for buck" solution provider is caching. Assuming that you had no problems *before* the slashdotting, that will solve your problem because it will cause Zope to need to serve far fewer requests, closer to the number of requests you normally get. And this is (I assume) the outcome that you actually want. I highly recommend setting up a caching proxy in front of Zope if this sort of load will be recurring. It's way faster and cheaper than trying to understand the problem deeply. ;-) Most commercial sites are developed using this principle, AFAICT.
butting in with questions :) with 2.3.x, RAM Cache Manager has no refresh/expire/max_age feature. you get that with 2.4.x and above only. so that might give users stale pages. i know, i've encountered it. i *really* think that there should be a RAM Cache Manager howto/best approach. i'd like to write it, but i need inormation. i'm currently caching my pages, but i'm not sure i'm doing it right. properties such as REQUEST variables (there's two forms available. 1- at the method's page, 2-at the cachemanager page), statistics (seems that the url does have some effect on the cache hit rates), etc. confirmation from product author/zc would really help. i guess that i should really look at the source ..., and i will. today. now, for putting squid in front of zope. how do others do this, when you have more than one server (using zeo). there's the virtual parameter in squid, but i don't get it. any body done this? my solution for now is putting in caching manager, and if *really* necessary adding in another zeo client. thanks
butting in with questions :) with 2.3.x, RAM Cache Manager has no refresh/expire/max_age feature. you get that with 2.4.x and above only. so that might give users stale pages. i know, i've encountered it.
Didn't someone make a RAM Cache Manager with a max age thing? I swear I remember seeing it posted on Zope.org.
i *really* think that there should be a RAM Cache Manager howto/best approach. i'd like to write it, but i need inormation. i'm currently caching my pages, but i'm not sure i'm doing it right. properties such as REQUEST variables (there's two forms available. 1- at the method's page, 2-at the cachemanager page), statistics (seems that the url does have some effect on the cache hit rates), etc. confirmation from product author/zc would really help. i guess that i should really look at the source ..., and i will. today.
The nuances of the cache manager baffle me as well, and I'll need to defer mostly to Shane, but I *think* that at least for the RAM cache the REQUEST variables listed are used as part of the key into the big "cache dict" where things are held. This means that, if, for example, you've got AUTHENTICATED_USER as a request variable, if the AUTHENTICATED_USER is "fred" when you cache one page, another user "bob" won't see fred's cached page (which might contain his name). However you *will* still be able to cache the page for fred.. the next time you ask for the page, a customized "fred" version will be served from cache. The more variables that are different and listed, the more RAM will be consumed for slightly different versions of the same page.
now, for putting squid in front of zope. how do others do this, when you > have more than one server (using zeo). there's the virtual parameter in squid, but i don't get it. any body done this?
There was recently a Zope-maillist posting by Sebas ICK that quoted Richard Barrett which goes in to some speculative detail about using Squid to front end for multiple Zopes. If it's accurate, it should really be cleaned up and turned into a howto.
my solution for now is putting in caching manager, and if *really* necessary adding in another zeo client.
I'd consider something like Squid to do the job of taking the load off your servers. It's a learning experience, but probably worth it for a site like yours. - C
Didn't someone make a RAM Cache Manager with a max age thing? I swear I remember seeing it posted on Zope.org.
yes. someone did. but i discovered it a bit late, as i've replaced the 2.3.x StandardCacheManager with 2.4.x :P all's working now, i hope
The nuances of the cache manager baffle me as well, and I'll need to defer mostly to Shane, but I *think* that at least for the RAM cache the REQUEST variables listed are used as part of the key into the big "cache dict" where things are held.
This means that, if, for example, you've got AUTHENTICATED_USER as a request variable, if the AUTHENTICATED_USER is "fred" when you cache one page, another user "bob" won't see fred's cached page (which might contain his name). However you *will* still be able to cache the page for fred.. the next time you ask for the page, a customized "fred" version will be served from cache. The more variables that are different and listed, the more RAM will be consumed for slightly different versions of the same page.
ok. there's two places where we can define this. the cache manager, and the methods cache tab. consider this: -we define REQUEST variables AUTHENTICATED_USER at the cache manager setting we dont define anything at the method cache tab -or we have somevar as the REQUEST variables in the method cache tab, and only AUTHENTICATED_USER in the cache manager setting. there's so many combinations that boggles the mind.
There was recently a Zope-maillist posting by Sebas ICK that quoted Richard Barrett which goes in to some speculative detail about using Squid to front end for multiple Zopes. If it's accurate, it should really be cleaned up and turned into a howto.
http://aspn.activestate.com/ASPN/Mail/Message/906749 i need to relook at this. what i have now: __ zeo client ----------- | : | lvs --------| : |-----zeo server |__zeo client------------| so with squid in the picture, i guess it'd be before lvs ..? thanks to sean, i know what i missed when trying squid with zeos. a redirector. (the squid virtual stuff pulls me of track :))
I'd consider something like Squid to do the job of taking the load off your servers. It's a learning experience, but probably worth it for a site like yours.
i'll look into this. thanks all
- C
there's so many combinations that boggles the mind.
Yeah. I'm afraid I know as much about it as you do at this point. ;-)
i'll look into this.
thanks all
Hopefully you let us know how it goes! - C
On Tue, 20 Nov 2001 14:40:25 -0600, "Michael Fraase" <mfraase@farces.com> wrote:
I don't run a proxy server or anything else in front of Zope.
This is almost certainly your problem.
And why does it fall over anyway? This just doesn't make any sense to me. I can see it getting slow and timing out, but giving up completely and just bailing? What's that about? Explain it to me like I'm an intelligent, non-technical friend. Thanks.
Zope's http implementation is *not* *robust* enough to be exposed to the raw internet. It has a number of serious, and fairly obvious denial-of-service vulnerabilities. Ive never seen Zope fall over without someone trying, but I wouldnt be suprised. Making it robust would involve a rewrite. Today the only way to make zope robust is to use a front-end proxy to sanitize the http conversation. If that front-end proxy also does some caching then thats a bonus. Toby Dickenson tdickenson@geminidataloggers.com
Zope's http implementation is *not* *robust* enough to be exposed to the raw internet. It has a number of serious, and fairly obvious denial-of-service vulnerabilities.
Toby, are these vulnerabilities in the collector? If not, and you've got the time, could you put them in there? - C
participants (14)
-
bak -
Chris McDonough -
Chris Withers -
Dieter Maurer -
Luke Tymowski -
marc lindahl -
Mario Valente -
Markus Schaber -
Michael Fraase -
Paul Zwarts -
Paz -
Richard Jones -
Thomas B. Passin -
Toby Dickenson