Tag Archives: host services

the battle…

the battle of other peoples’ computer started before i left for the fair.

about a week before the fair, my (now former) host provider decided to move the machine on which all of my web sites live from californa to florida. in the move, they somehow freaked out my VPN, because it started giving me an error that said “socket operation timed out” whenever i tried to send email. i could receive email just fine, but no sending.

and, because of the fact that the VPN isn’t SUPPOSED to interfere with normal network functions, i didn’t suspect that was the cause.

then my clients started reporting intermittent email failures. so i contacted my (now former) host provider and inquired about answers to what you would expect to be their problem (since they are, now “were”, my host provider).

as the week progressed, there was also concern that my scheduled web maintenance (which i wanted to have happen before the fair) was not possible, because of the ongoing problems with the web server, combined with the host-provider-support-drones who couldn’t speak english, answered (or not) only one question at a time, and who persisted in giving me scripted instructions for how (not) to fix the problem on a windoesn’t machine.

fortunately, most of my clients were also at the fair, and couldn’t access their email anyway. i gave up after being there one day, and only getting spotty WIFI coverage on my personal hotspot.

when i got home, much to my surprise, the ticket i had opened to fix the problem had been closed, but there had been no “fixing”… they closed the ticket because i hadn’t responded to it in three days, not because they had fixed the problem. almost as though they thought, if i wasn’t there to harrass them, there wasn’t a problem that they needed to work on.

i reprimanded them, and proceded to deal with my, now worsening, email problem. macque couldn’t access the list, monique’s email was spotty at best, a number of people said they had “given up” on getting messages to the list, and i STILL could only receive email, but not send.

i tried to deal with the now former web host for another week, getting less and less actual information, and more and more frustrated with my inability to communicate in anything like english (which strikes me as somewhat odd, given that it’s a british company, located in london). finally i had enough, and switched host providers…

AGAIN… 8/

it’s been about five years since i switched to EZPZ and ended my spree of switching host providers about every 6 months for a couple of years, and, at first, they were cheap AND excellent host providers… but in the past couple of years, i guess they must have outsourced their support, or something, because suddenly their tech support got CONSIDERABLY worse, and their language skills steadily decreased to the point where i would have to, DELIBERATELY avoid asking more than one question per ticket, in order to get anything close to a complete answer…

so, i switched host providers about a week ago. they quickly diagnosed my VPN problem, and got monique’s email working, sort of… but i forgot that i don’t have complete control over all my domains (Rev. Guido DeLuxe is actually the owner of two of them), and, instead of getting better, monique’s email problem actually got worse…

and the new host provider — ionblade — had conveniently forgotten to add “mailing lists” to my cpanel, so the two mailing lists i manage were completely offline…

and, meanwhile, i had to explain to macque that sending to “official-list-owner” sent mail to me, while sending to “official-list” sent mail to the list, and, because of the fact that i couldn’t send email, i couldn’t actually do something sensible, like forward his message… and even if i had been able to forward his message, i would have gotten a bounce also, because the FUCKING EMAIL LIST WAS DOWN!!

i actually helped them diagnose the VPN problem: i learned that i could access webmail at hybridelephant, but that it forwarded me to their server, and i could send from that, which i did for about a week. then the VPN updated, and quit working all together on my tablet, which meant that i had to do the “run-around” on my linux box when i absolutely HAD to respond to someone. basically, i took the message to which i had to respond, responded in my email client, copied the message, switched to a browser with webmail open, hit “compose new message” and pasted the message into that, in order to be able to send. when i couldn’t do that, either because my computer was down for the night, or i didn’t have access to it, i would use the browser on my tablet to respond. at some point within the past few days, it dawned on me that if i could send from webmail from port 80, even though my VPN was up, maybe the VPN was getting in the way, so i tried disabling it, and — lo and behold — the message went through. i enabled the VPN again, tried sending a message, and “socket operation timed out”. disabling the VPN made the message go through almost immediately.

so, i opened TWO tickets at the VPN provider, one about my tablet, and one to find out if there is some way to un-block my SMTP service on port 465. i (FINALLY) convinced the (new) host provider to re-add “mailing lists” to my cpanel, and, like magic, i suddenly had email lists that work (i think) again.

i (still) have two open tickets with the VPN provider, but my impression is that if they’ve got a bug that affects the connection of my tablet, it has probably struck a lot of people with similar tablets, so they’ve got a lot to deal with.

and i’ve still got web updates to get to, but the web update guy says that they can wait for a month or so, at which point there will probably be more anyway…

why is this happening?

several of my web hosting clients have been having trouble with their email not getting through, recently. i went through a frighteningly similar situation last year, which ended up with my moving my web sites, lock stock and CGI scripts, to another server with a different IP address… same host provider, just a different IP address. the reason was that, for some unexplained reason, the IP address kept getting listed in the CBL… which means that someone using that shared-hosting machine (not me, or any of my clients) was using it to sent spam.

the same thing has happened again. i complained, they suspended someone, it didn’t fix the problem, they told me that it takes time to propagate, i waited, it didn’t fix the problem, i complained again, they told me that it was an exim misconfiguration, that didn’t fix the problem, i complained again, they told me that they had written to the CBL to get answers…

well, guess what?

copied from the CBL home page: “There is no supporting data or “evidence” file available for any given listing, and no mechanism to ask why any given listing took place.”

so i wrote back to them, and said “whatever is causing the IP to be relisted hasn’t gone away, and won’t go away with you guys waiting for a response that won’t come.”

whereupon they closed the ticket with no further response.

maybe it’s time to start looking for a cheap VPS… 😐

a "smart" spammer??

i may have worked out why the mailing lists that i administer have been not delivering messages to all of the registered members of those lists recently. it was because of a “smart” spammer.

if you recall, Rule #3 of The Rules of Spam states that Spammers are stooOOpid, so how could a “smart” spammer even exist in the first place?

in this case, it’s a matter of degrees. you can be smarter than the ordinary spammer, but if you’re a spammer yourself, you’re still pretty stooOOpid compared to those of us who aren’t spammers.

the “smart” spammer shared my IP address, and was very definitely sending out spam, but he was not flooding the net with spam, like most spammers do, he was trickling spam out at a rate that was just below the point at which my host providers’ spam alarms would indicate. thus, while other, more brash spammers got the mallet immediately, this “smart” spammer continued to share my IP address for several months before he got the mallet.

of course, this was also complicated by the fact that, for several weeks, my host providers’ answer to all of these problems i was having, was to delist my IP address from the CBL without actually attempting to discover why that address was listed at the CBL to begin with. which is where persistence comes into play. instead of accepting that “the host provider knows best” and closing the work ticket the first time they requested delisting, i actually checked CBL myself, to determine whether or not what the host provider had told me was true, and when it turned out that my IP address had been relisted, i called them on it… and then, i called them on it again, when they told me that everything was the way it should be… and then, i called them on it again… and again… and AGAIN! 😐

i think they finally got the idea, because as of 10:00 this morning, my IP address had not been listed at the CBL for 24 hours, which is the longest time it has not been listed in 2 months.

let’s keep it that way, okay?

nameservers

my domain, hybridelephant.com has ns5/ns6/ns7/ns8.thewebhostserver.com as its nameservers: that is, when you go to any page on hybridelephant.com your request is actually handled by one of those four servers. the web site (and all of my clients’ web sites) actually lives on a server called us8.la.thewebhostserver.com.

because of the fact that i design and host web sites for the community, i have registered two nameservers of my own, ns1.hybridelephant.com and ns2.hybridelephant.com, and pointed them at two of my nameservers. they’re not “real” nameservers, but they work just as well, as long as they’re pointed at my real nameservers.

from time to time, i check IntoDNS to make sure internet knows where to find me. this afternoon, i checked and discovered a whole bunch of error messages where, when i checked the last time (a few months ago) everything was copascetic…

it’s generally been my impression that if I don’t change my DNS settings, they DON’T change unless something is SERIOUSLY wrong, so naturally i freaked out… just a little, but it was definitely something that made me sit up and take notice.

after some wangling around, it was determined that it was because i had hybridelephant.com’s nameservers set to hybridelephant.com. it works, it just throws errors all over the place. the solution, i come to discover, is to create “real” nameservers. to do that, i would have to “upgrade” my account to a “reseller” account. currently i have a “shared hosting” account. if i upgrade, in order to be able to afford the account, i would have to drop from 30gb of space (which is the “master”, or largest shared hosting account avaiable) to 5gb of space (which is the “starter” or smallest reseller account available), and that would still be more than twice what i am currently paying. the “up” side is that i would get to have “real” custom nameservers, and my clients would all get their own cPanel account (which they would have not the first clue about how to use)…

of course, the “starter” account comes with a lot of other stuff, like web templates and legal documents, and a free WHMCS license — that i don’t, can’t or won’t use — all for more than twice as much as i am currently paying for 30gb of space.

the only thing i would use is the “unlimited SSL certificates” but in fact, i only really need one, and i’m fairly sure there are other ways to get a SSL certificate…

at this point, i can deal with the intoDNS errors… it works, i get a pointer to my nameservers for my clients who don’t understand, and i get FIVE TIMES AS MUCH disk space, for more than HALF the money…

grmph!

it’s getting really close to opening night for the panto, and the fremont players domain has developed a DNS hiccup that i haven’t been able to figure out. apparently, “ticking the box that says “Use the nameservers specified at the Domain’s Registrar (ignore locally specified nameservers)” on the “Create a New Account” page” isn’t part of the deal, but i haven’t used a “Create a New Account” page on fremontplayers.com for almost a year, and certainly not before i switched to the new host server, approximately 5 months ago… even more bizarre is the fact that it was working just fine yesterday.

not only that, but the webdavs web disk applet that i use has suddenly decided not to work: instead of connecting me to the machine in california on which hybrid elephant lives, it hangs up and won’t go any further on my mac, and on the linux box, everything loads, but when i try to copy or move files, it says “the disk ‘hybridelephant.com’ doesn’t exist”. i can move things around with ftp and ssh over a text-based connection, but i’ve really come to appreciate the reality of doing all that technical web stuff with a graphical user interface…