Email woes

[rkstar]

In the past six weeks or so, my site has been down for hours at a time... 3 times so far. Is this going to become a regular occurance? Don't they offer some kind of uptime guarantee? And the email server all of a sudden started bouncing back more than 50% (in the thousands) of my emails from my newsletter. Then after tech support says they fixed the problem by "updating the exim", it still didn't work. So I tell them it was still broken, and they tell me they fixed the problem by "updating the exim". What are the chances it worked this time?

feeling rather frustrated,

-dan goodspeed
www.rkstar.com

[Vin DSL]

Nope! Going pretty smooth, since I changed neighborhoods...

[rkstar]

neighborhoods?

[Zacrifice]

I haven't had problems with my site however a few people at support seem to be a little blind or absolutely stupid! (Thats the first bad thing I have ever said about them too)
I sugest sending another support ticket detailing everything again!
Then just stay on there backs!

Unless it's perhaps you
Has your ISp recently made any changes?
do you have monitoring for your site? (externaly to check uptime!)

[Rye Seronie Oh]

I've been hosted on JPC for close to 2 years I think and the last time there was a service issue, it was because SBC DSL for the Houston area picked up a brand new IP Block and that range wasn't allowed on the servers here or several other places on the net because they were so new. JPC fixed it immediately though!

I've never found anyone in the support team that is 'blind' as you say.

Anyhow, perhaps you should run a trace to your site and see where the failure is so we have a better picture of why you're unable to access your site. Are you having difficulty accessing any other sites on the web at all? Are you on the Gigadeal/Longhorn or something like a VPS/Dedicated?

[JPC-Masood]

If you are sending over 1000 mails per hour, and your emails get rejected by server, then updating exim will not fix it. You will need to put a delay in your bulk mailing so that it does not overload the shared server. Otherwise the server will just get blacklisted due to many of the form mail header injection exploits that exist.

If even after putting the delay, you are facing problem with email, please send in a ticket with detailed steps to reproduce the problem.

[Vin DSL]

neighborhoods?

Servers...

I do a lot of coding in the 'wee hours' and the server I was on for 2 years started having problems, nightly, when their backups kicked in. 'We' (Tech Support and myself) suffered through this for 5 months, before I requested a server change. Since they transferred my account to another server, everything has been fine -- knock on wood...

[Connie]

all of a sudden started bouncing back more than 50% (in the thousands) of my emails from my newsletter.

Sorry I have to ask. Do you really have that many subscribers? If so I would think you need to be on a dedicated server rather than a shared server.

I have sites hosted at Jag for years and have never experienced the problems you seem to be having.

[gerilya]

2 quick questions:
- do you mean your whole website was down for hours?
- How do you send newsletters? Are you using your hosting account's SMTP or you use newsletter software installed on your account?

[gerilya]

BTW, I suggest you ask support to elaborate on the steps they took to resolve this problem. For example, ask name of a person performing the update and what exim version was previously installed on that server and what version they upgraded it to

This won't make the problem disappear, but it will ensure the problem is taken more seriously next time.

We've had quite a few rather unpleasant experiences where support was lacking a knowledge to resolve quite simple problems. Not only that, they also lied about how they tried to fix them.

[JPC-Greg]

We've had quite a few rather unpleasant experiences where support was lacking a knowledge to resolve quite simple problems. Not only that, they also lied about how they tried to fix them.

Here with us?

[rkstar]

I'll see if I can answer everyone here -

Zacrifice- It's definitely not my ISP. When my site goes down the emails and IMs come flooding in to let me know. The users are my "monitors"

Rye Seronie Oh- It happens so frequently lately that I don't bother with a Traceroute, and the tech support people acknowledge "it's a temporary problem", so it's definitely a Jaguar problem. I'm currently on the $10/month plan whichever that is.

Masood- Yeah, they just told me about the 1,000/hour problem. Which is weird because i'm actually sending less per hour now than i have done before, and it's weird that now is when the problem just started happening (after years of working fine). I'll add in a 4 second pause between each email to guarantee that no more than 900 emails can be sent per hour, but it still bothers me that some people who are in the first thousand of the email list are getting bounced back... kinda indicating that it's not the number of emails I'm sending... as well as occasional emails after the first 1,000 (it's really hard for me to trace who's who, so i don't have exact stats).

Vin- I have been thinking to upgrade servers to the $30/month plan when i release my new version of the site, which i'm expecting to be pretty big... like 10x the number of daily hits, and at least twice the subscribers. Though I was kind of taken aback by one of the times the server went down, Jaguar's suggested solution was for me to upgrade the server. It kind of didn't seem right that their suggestion to get my website back online was to pay them more than 3 times what my current account is. Especially when I only use 5-10GB of bandwidth/month, when the account offers 180GB. Either way, I'll probably still upgrade once the new site is ready.

clssam- I have about 3,000 subscribers (and getting more each day). But like I said, I don't use that much bandwidth, and don't really do anything that requires a super-fast computer. I write all my perl scripts carefully to be as efficient as possible for the site, rather than using cookie-cutter scripts. I really haven't had any problems with Jag til the last couple of months.

gerilya- Yes my whole website was down for hours. You go to the site and it just says "Server Not Found"... and people all over were reporting the problem to me. I send the newsletters with my own perl script, each user gets a customized newsletter to their preferences (the whole Web 2.0 thing) so a generic newsletter program wouldn't work. I send each message using sendmail. And yes, I believe it is my hosting account's SMTP.

Thanks for all your comments/suggestions!

-dan goodspeed
www.rkstar.com

[gerilya]

Yes my whole website was down for hours. You go to the site and it just says "Server Not Found"... and people all over were reporting the problem to me.

When you notice the server is not found, make sure first it is not a problem of your DNS: try connecting to the server using IP address, i.e., http://321.321.321.321/~username

If you have SSH access to the server, try it too.
If you can't SSH to your account using iP address - it is a server-wide problem that affects all clients on that server and you should open support ticket asap. Don't wait for the site to get back online - it doesn't work that way here as there is no proactive service monitoring.

I send the newsletters with my own perl script, each user gets a customized newsletter to their preferences (the whole Web 2.0 thing) so a generic newsletter program wouldn't work. I send each message using sendmail. And yes, I believe it is my hosting account's SMTP.

Need more info on this one still: there are 3 ways for your script to send mails:

1. connecting mailserver of the recipient directly
2. connecting your server's local SMTP server
3. writing mail content into sendmail program via pipe

I assume you use the 3rd one as it's easiest. In that case you actually submit the mails for delivery rather than sending them immediately. On properly configured servers sendmail/exim/postfix/whatever should automatically limit a number of outgoing mails it sends per some time interval.
There is no need for you to insert delays into the script as it won't change anything.

How the bounces look like?

[gerilya]

If you are sending over 1000 mails per hour, and your emails get rejected by server, then updating exim will not fix it. You will need to put a delay in your bulk mailing so that it does not overload the shared server. Otherwise the server will just get blacklisted due to many of the form mail header injection exploits that exist.

I don't get that part: who will blacklist the server?

[rkstar]

gerilya, i believe you're right in that i am sending the mail via pipe... this is kinda what my source code looks like for the send newsletter script-

open(MAIL, "|/usr/sbin/sendmail -t -f rkstar\@atom.nocdirect.com");
print MAIL "To: $userinfo{email}\n";
print MAIL "From: editor\@rkstar.com\n";
print MAIL "Reply-To: editor\@rkstar.com\n";
print MAIL "Errors-To: rkstar\@atom.nocdirect.com\n";
print MAIL "Subject: $sSubject\n\n";
print MAIL $message;
close(MAIL);

And with that I'm getting thousands of returned emails all stating the same error message:

Envelope-to: rkstar@atom.nocdirect.com
Delivery-date: Fri, 17 Mar 2006 18:07:50 -0600
Received: from mailnull by atom.nocdirect.com with local (Exim 4.52)
id 1FKOz0-0004Ii-LA
for rkstar@atom.nocdirect.com; Fri, 17 Mar 2006 18:07:50 -0600
X-Failed-Recipients: valid-email@hotmail.com
Auto-Submitted: auto-generated
From: Mail Delivery System <Mailer-Daemon@atom.nocdirect.com>
To: rkstar@atom.nocdirect.com
Subject: Mail delivery failed: returning message to sender
Message-Id: <E1FKOz0-0004Ii-LA@atom.nocdirect.com>
Date: Fri, 17 Mar 2006 18:07:50 -0600

This message was created automatically by mail delivery software.

A message that you sent could not be delivered to one or more of its
recipients. This is a permanent error. The following address(es) failed:

valid-email@hotmail.com
unrouteable mail domain "hotmail.com"

-----

it might also be of interest i just noticed the EXIM version was 4.52 before the first time they said they updated it, still 4.52 after they said they updated it, and it's still 4.52 now that they say they've updated it again. According to the exim website, the latest version is 4.6.0.

And thanks for the suggestion about connecting via the server ip. if it goes down again, i'll try that. Though I really do think it's server related. The last time it went down, it first started by all perl scripts on the server returning "internal server error" for no reason, so i submitted a trouble ticket about it. 15 minutes later the whole site stops responding, i figure they were just restarting the server... but it was at least 5 hours before it came back online.

Any way... I have to send out this week's issue now, fingers crossed it actually reaches all the subscribers... but somehow i doubt it will work.... for the 5th week in a row.