I had just started a small forum site <<< removed >>> and it quickly bogged down and then became unreachable ("connection refused") by nighttime. The next morning it was reachable, but it is slow. My host asked for a tracert output when I asked about the problem, then said it was due to my isp because of a timeout. From what I see, though, the timeout occurs at a router very close to their end, not near mine. Furthermore, I see timeouts when I run traceroutes from other servers on the net, as well as from another ISP in town (using DSL instead of cable modem). These timeouts aren't in exactly the same place, but they are, again, near my sites origins.
On the other hand, in my town it's mostly the people who are also using cable modems that see the slow behavior, not so much the DSL customers, although they are theroetically similar speeds (I'm in Mexico, btw).
So I'm a little unsure whom to put pressure upon. There are only a couple dozen users, so it's not a heavy load.
Anybody have any insight, suggest any other diagnostic means I might use?
Thanks, and hope this is appropriate since it's only indirectly about webhosting per se.
Last edited by SoftWareRevue; 06-23-2005 at 02:34 PM.
Just to let you know the site appears to be loading quite quickly at my end here in the UK, and I am using NTL cable internet. This could indeed be an issue with your ISP and not actually your host.
What I would suggest is that you obtain a free account with SiteUptime.com or KaneAtInternetSeer and setup remote monitoring of your website, whenever your site goes offline you will be notified, if you use SiteUptime.com they use remote locations from across the globe to monitor your site, you can then present your host with the overall downtime reports for them to investigate further.
The above should help you pinpoint whether this is an ISP or Host specific issue without any problems. Best of luck, keep us posted.
█ Tahir Ahmed █NetspaceOnline.net - Reliable Personal Web Site Hosting & Business Web Hosting Solutions! █NSDomains.net - Our Dedicated Domain Registration & Management Portal! █ • Reseller Hosting Solutions • Linux Plesk 8 Control Panel • Provding Quality Hosting Since 2001
Check your MaxClients in the your apache config, make sure its around 256 or so (if your hardware is decent). It could just have a max number of clients reached. Also whats the load on the box during these times?? Have you bee remotely connected when its occured?
Head on over to siteuptime.com and sign up for a free account. Use that to monitor your site for a week, and then forward the results to your host. It sounds like a server issue. ask your host for a resolution.
GenerousWebHosting.com Home of Generous George the Red Monster. 24/7 Live Chat Support - 99.9% Uptime - 30 Day Money Back Policy
Yes, I have set up at siteuptime.com as suggested earlier, and it shows (checking every 1/2 hour) 100% uptime and .06 sec. response time from London (the site is hosted in Texas apparenly).
I saw a techie from the cable company and showed him a traceroute (or maybe it was a pathping) which I executed at their office, and he said he'd report it and have things checked further down the line, if I understood correctly (my Spanish is limited). Meanwhile any DSL machine I've tried (e.g. at cybercafes) seem very speedy by comparison.
I don't have access to the apache config, I'm on shared hosting.
And, oh, sorry about the awful formatting of the traceroute above!
Appreciate everyone's input, and will post what the techs come back with (could take a few days here in Mexico...)
In looking back over this thread, I see that what I posted wasn't as clear as it should have been. Specifically, I consistently got a timeout at a specific node:
17 100 ms 107 ms 108 ms dcr2-sto o-5-0-0.LosAngeles.savvis.net [220.127.116.11]
18 * * * Request timed out.
19 148 ms * * bcr2-so-0-0-0-500.Dallas.savvis.net [18.104.22.168]
That is, there appeared to be a problem with packets traveling between LA and Dallas. As my ISP wasn't forthcoming in trying to fix the problem (and remember other ISPs weren't experiencing this problem), I actually contacted "savvis.net", who is a big network provider, telling them that I thought they might have a problem between those two nodes. They responded (within 1/2 hour of my email!) saying that as far as they could tell, their networks were operating normally! I went back to my ISP one more time (this time with a Mexican friend who was fluent with both Spanish AND geek-speak!) to try one more time to pursuade them to work on the problem. Simultaneously, my host (YRHost) took the initiative and went ahead and re-located my site to a different server. After 24 hours (nameserver changes) my site worked just fine! At the same time, though, my old IP address traceroute now doesn't indicate a timeout, either (although the route is still identical), so perhaps BOTH parties (host AND ISP) finally did what was necessary!
The funny thing about the ISP is that their route goes up to San Jose and THEN down to Texas even though the packets originate down here near Mexico City!
Meanwhile, I've had the site monitored by siteup.com (thanks for that suggestion, folks) and it's been just fine (from London, England!) all along...
Thanks for all your input. Just wanted to not leave you all hanging! You can start breathing now...