Answers below..... -----Original Message----- From: John Wright [mailto:[EMAIL PROTECTED] Sent: Friday, April 02, 2004 11:05 AM To: inFusion Support List Subject: Re: [iMS] RAMDisk results
Thanks David. Your hardware/os specs sound pretty typical. You mention two interesting things though - Tokens and DNS servers. - How do tokens affect performance of the CFX tag? Without tokens the POST module can create multiple mail files for each control file. Is this correct Howie? Is it better to loop over the CFX tag and pass single tokens in or use the QTOKEN functionality? We found out very quickly not to use any loops. This made the process crawl and crash. Or is it better to not use tokens at all (which implies looping over the CFX tag for each unique email)? Passing no variables is the fastest however was acceptable for tracking. A unique ID for each email was passed to track bounces, opens, clicks, opt outs... All this processing was done on another subnet. Are you running the process the creates the emails with the CFX tag on the same server that is sending with the POST module? Yes cfx tag and POST mod. was on the same server. - You have 10 DNS servers configured! I only have two (maintained by my ISP), but I have my MX cache set to 10,000. Is this a significant bottleneck? If I recall correctly, Howie had me set MX cache to 500,000. DNS can be a significant bottleneck because POST module has to resolve MX domain names to IP addresses before sending. - Was your rate of 300K/hour observed in the real-world (vs. a test setup)? This was real world. Our client FTPed a flat file of 15 Million opt in email addresses. The goal was to get out email to the whole database as quickly as possible. John ----- Original Message ----- From: "DGo 800-826-7763 x117" <[EMAIL PROTECTED]> To: "inFusion Support List" <[EMAIL PROTECTED]> Sent: Friday, April 02, 2004 9:15 AM Subject: RE: [iMS] RAMDisk results We did not use ram disk software. Howie never brought it up, so we did not experiment with it. Here are the six factors determining performance of the POST module: 1. Size of the email - (assume 10k html for calculations.) 2. Tokens - Passing variables into the email. 3. MX Servers on the other end. 4. Size of the Queue 5. Server Hardware and Software Setup 6. Bandwidth I would ask Howie to put these in order of importance and anyone on the list to provide input. Server hardware was dual processor, 2 Gig. RAM, Adaptec wide, Seagate 10k rpm Cheetah 18 gig drive. Server software was Win2k Advance Server (don't know if this made a difference), with one 2 gig. Partition for OS , 16 for database, Ims Out and Ims logs folders. No RAID setup. Applications were CF Ent. 5, MSSQL 7, IMS POST. (Howie said it was better to run the SQL on the same machine. Don't know why?) Lan was 100 meg. Switch to Cisco 4500 and WAN was DS-3 a few hops from Uunet's backbone. We had two Linux DNS servers on the same subnet and set 8 more in IMS Config. on tier 1 backbone. More later.. Regards to all, DGo ==^======================================================= This list server is Powered by iMS "The Swiss Army Knife of Mail Servers" -------------------------------------------------------------------------------------- This list is provided as a free service. Although we will try to address issues in a timely manner, support via this list is not guaranteed. If you require expedited support then a support contract is required. Support may be purchased from http://www.coolfusion.com/commerce. Details regarding support options may be reviewed at: http://www.coolfusion.com/SupportOptions.cfm -------------------------------------------------------------------------------------- To leave this list please complete the form at http://www.coolfusion.com/Support/ Need an iMS Developer license? Sign up for a free license here: http://www.coolfusion.com/Developers/ List archives: http://www.mail-archive.com/infusion-email%40eoscape.com/ Note: You are subscribed as [EMAIL PROTECTED] ==^=======================================================
