The reason I ask is I'm unsure if the machine is having issues and I'd like to see if it's a machine issue or what.

Also, I sure hope there aren't any issues... It's a RAID5 on a gigabit uplink.

-- Kate Gerry OC3 Networks & Web Solutions 530 W 6th Street #901 Los Angeles, CA 90014 kate@oc3networks.com

-----Original Message----- From: centos-mirror-bounces@centos.org [mailto:centos-mirror-bounces@centos.org] On Behalf Of mirror-maintainer@mirror.averse.net Sent: Saturday, February 07, 2009 15:19 To: Mailing list for CentOS mirrors. Subject: Re: [CentOS-mirror] Good stress tester...

On Sat, 7 Feb 2009, Kate Gerry wrote:

I've been setting up a mirror and I would like to know if anybody has any good ideas on how to stress test it without being live?

I'd like to be able to simulate a lot of requests from either one or multiple IPs (I know I'd have to run it on more than one system)

Please let me know as I'm not sure if it's working correctly as of

yet!

IMO, real world traffic for static files (ie, normal mirror servers) are

unlikely to stress mirrors much. The bottleneck is likely to be available bandwidth, I/O, or possibly RAM/available threads/file descriptors to serve requests.

Unless you're just trying to see at what the limit really is at which your server will fail... _______________________________________________ CentOS-mirror mailing list CentOS-mirror@centos.org http://lists.centos.org/mailman/listinfo/centos-mirror

Hi there,

what makes you believe the machine is having issues when it has not been made public and served any load yet?

There is two software suites I know of to do automated stress test on applications or websites as in your scenario

http://www-01.ibm.com/software/awdtools/tester/robot/index.html http://www.automatedqa.com/products/testcomplete/

But before you invest in such a license and especially time in learning how to use the software, you are probably better off reading logs pointing to a hardware failure or just doing away with the machine. A mirror is an ideal environment to "just see when it fails" and have some network load normal websites don't see. It won't hurt anybody, the built in failover in yum will just pick another mirror if the mirror machine goes down or stops serving requests due to overload.

Sorry for being nearly cynical, but your e-mail address points to a network company and I suppose there is some people who back up the offerings with actual know-how do diagnose issues once they happen. You needn't worry about the design of the mirror as it has been a proven concept working well for many of us. Be confident, roll it out and let it happen, whatever it be.

Florian

----- Original Message ----- From: "Kate Gerry" kate@oc3networks.com To: "Mailing list for CentOS mirrors." centos-mirror@centos.org Sent: Sunday, February 08, 2009 4:40 AM Subject: Re: [CentOS-mirror] Good stress tester...

The reason I ask is I'm unsure if the machine is having issues and I'd like to see if it's a machine issue or what.

Also, I sure hope there aren't any issues... It's a RAID5 on a gigabit uplink.

-- Kate Gerry OC3 Networks & Web Solutions 530 W 6th Street #901 Los Angeles, CA 90014 kate@oc3networks.com

-----Original Message----- From: centos-mirror-bounces@centos.org [mailto:centos-mirror-bounces@centos.org] On Behalf Of mirror-maintainer@mirror.averse.net Sent: Saturday, February 07, 2009 15:19 To: Mailing list for CentOS mirrors. Subject: Re: [CentOS-mirror] Good stress tester...

On Sat, 7 Feb 2009, Kate Gerry wrote:

...

I've been setting up a mirror and I would like to know if anybody has any good ideas on how to stress test it without being live?

I'd like to be able to simulate a lot of requests from either one or multiple IPs (I know I'd have to run it on more than one system)

Please let me know as I'm not sure if it's working correctly as of

yet!

IMO, real world traffic for static files (ie, normal mirror servers) are

unlikely to stress mirrors much. The bottleneck is likely to be available bandwidth, I/O, or possibly RAM/available threads/file descriptors to serve requests.

Unless you're just trying to see at what the limit really is at which your server will fail... _______________________________________________ CentOS-mirror mailing list CentOS-mirror@centos.org http://lists.centos.org/mailman/listinfo/centos-mirror _______________________________________________ CentOS-mirror mailing list CentOS-mirror@centos.org http://lists.centos.org/mailman/listinfo/centos-mirror

Silly enough, the reason why I ask is because I'm not 100% on my RAID + RAID controller. I want to ensure that it's stable and not dropping out/etc. And usually, a real 'stress' will show you. While the redundancy of the yum setup is fine... I'd rather not have the attitude of "If it fails it's not a big deal", I want to feel that if I set it up, I don't need to worry about whether it's still up and working at its peak every 5 minutes.

Server info:

FreeBSD 7.1 i386 Dell 2950 /w 4-bay slot raid Xeon 3.0Ghz (Dual Core + HT) 2GB DDR2 Broadcom NetXtreme II BCM5708 1000Base-T attached to Cisco SRW2024 /w fiber uplink 1x 146GB SAS Drive (OS) (I wanted to do a RAID1 on this but since I grabbed a chassis with only 4 ports instead of 6, I was limited) 3x 1TB SATA2 Drives (RAID5 /w repos on it)

I just thought I'd set this up 'for the lulz' and so far it's worked okay. However, I haven't been sure if there have been problems on the server-side, network-side, or just inside my head. There aren't any errors listed in any logs... But logs aren't always everything.

-- Kate Gerry OC3 Networks & Web Solutions 530 W 6th Street #901 Los Angeles, CA 90014 kate@oc3networks.com