For any of the other special snowflakes out there had loaded up a new instance of clear and then scratched their heads about why it couldn't register...well it might be because it didn't set itself up a default route.
Checking the routing table by typing 'route' showed only the local subnet with gateway of 0.0.0.0; mask etc etc.
If this happens, one might like to try dropping to a text console (Alt-F2), logging in, and typing the following:
or eth1 or whatever. x.x.x.x is obviously whatever the thing is supposed to be talking to whenever it wants to get outside the local net.
Perhaps I'm just missing the obvious in the forums but I will risk a pair of stupid questions.
1) Is the ClearVM setup more similar to XenServer in that it installs a hypervisor only with extremely minimal Dom0 host base, or is it adding the KVM functionality and management tools to a standard ClearOS setup? Disk management and planning are required for either scenario, but the infographics don't really say anything technically useful.
2) Because management tools for Xen suck, I would like to migrate off of XenServer. I've been contemplating using a bare linux distro and loading up Xen on that or switching to KVM. Can ClearVM fairly easily migrate in a Xen vm and run it in KVM/ClearVM?
I was reviewing the release notes for 7.2, I'm curious if the IPv6 DNS item listed as bug 6361 was the original problem here. Maybe one of these weekends I'll spin up a new vm and test it. I do have IPv6 available, and it is generally much faster than IPv4 for some reason or other. I really do <3 clear.
Just did a full reload. Was very careful to slowly install things one at a time. Squid. Gateway Antivirus. Dans. As soon as dans is enabled it pretty much wets the bed. I'll keep poking into things later on when the weekend comes.
I did a diff between my old clearos6/etc/dansguardian-av and clearos7/etc/dansguardian-av.
Interesting bits between dansguardian.conf:
increasing log level on new machine
< loglevel = 1
> loglevel = 3
< logexceptionhits = 0
> logexceptionhits = 2
< weightedphrasemode = 0
> weightedphrasemode = 1
< filecachedir = '/tmp'
> filecachedir = '/var/cache/dansguardian'
< contentscannertimeout = 20
> contentscannertimeout = 60
< logconnectionhandlingerrors = off
> logconnectionhandlingerrors = on
I just reinstalled squid and dans. I browsed a bit without either, then with squid, then with dans added. Performance tanked with dans as before. The squid log doesn't really show anything interesting. I did something along the lines of
against the /var/log/dansguardian/access.log
The following very clearly shows that dansguardian spends the next 4 minutes chewing itself through a simple local newspaper webpage. It takes less than 4 seconds to load with dansguardian off, and I've always considered dans to be completely transparent and usually forget it is there. I am super curious to know why it is suddenly tanking.
Been running clear since the clarkconnect days....never seen dansguardian perform so poorly as 7.1 with dans version 2.1.7-1.
Freshly loaded test machine pretty much unusable. I only really use it for virus scanning of material, so I disable all the lists. Set it up no_cache deny all just like before since it really doesn't take much at all to outstrip read/write speed of a spinning disk. Out of the box it was so unusable I thought it was poor dns, but that is definitely not the issue. Turning off dans, and running with just the squid (restarted) usually results in strange page errors saying such and such subdirectory of the page was unavailable. Removed dans and squid and everything was perfect. Reinstalled squid and it was still perfect. Installed Dans and all is strange again. Some pages load alright, others just sit and spin for a minute or two before loading. It is just me testing it, on the exact same hardware that it had before with no issues. Just plain strange.
I'll just run it with dans uninstalled until I see something useful pop out. I'm guessing it is some sort of goofy permissions or self resolution problem.