Out of the blue I can't connect to neither my site (www.olegdulin.com), or via ssh to my linode or host31.linode.com .. Ping them works fine, it seems. I also rebooted my linode to see if that does it.

No luck..

Any thoughts ?

Hello,

As I replied in your ticket when you opened it, I'm not having any of the problems you described. I can access the host, your website, and get a response from your ssh server.

Send me a traceroute, maybe it's a network problem between you and the data center...

-Chris

I am also on host31 and have recently had similar recurring problems connecting to my sites and through ssh (but ping works fine). Usually, if I wait for a little while, I can connect again. This has happened a few times.

Caker has informed me that my Linode was using excessive swap I/O, causing the host to slow down. It probably is due to my Linode.

Sorry for the inconvenience caused.

I'm now migrating to another VPS. I've since redirected all requests to Apache to my new server so loads should be reduced. Moving mail services now.

I have no idea what's causing all that swap usage. I've stopped backups and cronjobs (except AWStats, which runs 6-hourly).

I hope it will be corrected soon, the server is annoyingly slow for a few days
slow = almost always unreachable

I monitor host31 pretty closely via my Linode; when a host is overloaded with heavily swapping Linodes, my Linode's load rises to somewhere between 5 and 20+. I've found that this normally happens anywhere from a couple of times a day to once every couple of days. My logging doesn't show any deviation from that pattern in the past week, so if you're having problems with host31, likely it is either your Linode (possibly being slow to respond because it's being I/O limited because your Linode is doing something I/O intensive like swapping too much?), or your Internet route to the data center. I think that host31 has been behaving reasonably well for the past week. You should expect that every once in a while, your Linode will be unresponsive, but typically not for more than half an hour or so.

Here's my Linode load for the past week:

Hi,

Thanks for the detailed response on this host being slow. I have just had an unresponsive host31 where I could not SSH to either my own SSH server or the linode console.

I have also found that there are some hosts that I can't reach from host31. For example www.ultimatecarpage.com is unreachable:

# traceroute www.ultimatecarpage.com
traceroute to ultimatecarpage.com (67.15.20.40), 30 hops max, 40 byte packets
1 li9-73.members.linode.com (67.18.176.73) 2976.579 ms !H 3007.046 ms !H 3002.058 ms !H

My Network settings on Gentoo Linux are as follows:

iface_eth0="67.18.176.73 broadcast 67.18.176.255 netmask 255.255.255.0"
alias_eth0="67.18.176.74"
gateway="eth0/67.18.176.1"

Here is a traceroute to my linode. I can ping the server, but traceroutes and ssh are all dead

laptop root # traceroute host31.linode.com -n
traceroute to host31.linode.com (67.18.92.13), 30 hops max, 40 byte packets
1 * * *
2 195.149.20.11 18.947 ms 19.627 ms 19.856 ms
3 213.208.106.194 19.941 ms 19.926 ms 19.958 ms
4 195.149.20.137 22.016 ms 19.956 ms 20.149 ms
5 64.21.115.61 23.670 ms 22.060 ms 21.915 ms
6 209.123.11.30 103.775 ms 107.841 ms 113.962 ms
7 206.223.115.72 109.769 ms 105.837 ms 105.917 ms
8 216.140.8.10 117.848 ms 110.027 ms 103.617 ms
9 216.140.8.161 105.976 ms 105.751 ms 102.196 ms
10 216.140.8.110 115.617 ms 135.930 ms 129.690 ms
11 216.140.17.102 129.959 ms 143.755 ms 130.021 ms
12 216.140.4.70 129.520 ms 125.893 ms 135.878 ms
13 67.99.43.234 139.750 ms 137.832 ms 138.065 ms
14 12.96.160.11 155.762 ms 137.731 ms 135.756 ms
15 67.18.92.13 133.838 ms 133.957 ms 133.773 ms

laptop root # traceroute -n 67.18.176.73
traceroute to 67.18.176.73 (67.18.176.73), 30 hops max, 40 byte packets
1 * * *
2 195.149.20.11 18.407 ms 23.773 ms 19.777 ms
3 213.208.106.194 19.950 ms 39.666 ms 23.733 ms
4 195.149.20.137 19.858 ms 19.580 ms 19.705 ms
5 64.21.115.61 21.663 ms 19.915 ms 19.812 ms
6 209.123.11.30 93.579 ms 97.715 ms 99.841 ms
7 206.223.115.72 97.587 ms 95.752 ms 96.034 ms
8 216.140.8.10 97.427 ms 95.768 ms 97.657 ms
9 216.140.8.165 95.690 ms 97.655 ms 95.925 ms
10 216.140.8.22 113.872 ms 111.357 ms 107.779 ms
11 216.140.16.29 127.854 ms 216.140.17.106 127.454 ms 216.140.16.29 125.938 ms
12 216.140.4.70 131.476 ms 216.140.4.226 127.684 ms 216.140.4.70 147.875 ms
13 67.99.43.234 145.433 ms 137.745 ms 153.917 ms
14 12.96.160.44 133.430 ms 133.998 ms 135.472 ms
15 ...
30 * * *

I don't think any of my network settings are wrong as they are straight from the memebrs area.

The only difference I can see between Host31 and my linode is the IP range the network interface is on. Has anyone else experienced this sort of problem on the 67.18.176.x network?

This sounds to me like you've got a netmask incorrect on one of your eth0 aliases. Check the output of "ifconfig" and verify all the netmasks are 255.255.255.0.

-Chris

bingo!, that's fixed it

I added the following to my /etc/conf.d/net file

+# NB: The next is only used for aliases.
+#
+# To add a custom netmask/broadcast address to created aliases,
+# uncomment and change accordingly. Leave commented to assign
+# defaults for that interface.
+#
+broadcast_eth0="67.18.176.255"
+netmask_eth0="255.255.255.0"


Once again Chris you've saved the day

Cheers.