New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
We leave traffic shaping OFF in Solus, as it seriously fubar's a lot of stuff with the amount of IPs we average per node.
Yeah but for the users getting 80-110MB/s (the majority) everyone running them will slow it down :P
The way iptables works with the node density/ip count density we have, you won't realistically pull 80MB/s inside a container. On the host node, yes. Inside a container, it averages between 40-60
Brief update:
We have resolved temporarily issues for all but 1 container that we are using to troubleshoot further. Believe we may have found the particular issue. Thank you for the reports and patience
Edit: Everything should be all good to go. If anyone else experiences similar issues, please open a support ticket.
not too bad
What was it, out of curiosity?
Atrato has been having issues as of late in JAX, causing some intermittent connectivity issues. We've disabled our Atrato uplink until they give us the green light on resolution. Some of the containers (was about 10 or so) didn't want to come back up to speed for some reason so had to do some poking and prodding to get them to behave normally. Not 100% sure that both are related, but all containers are performing normally again.
Overall a very small subset of clients impacted by network speed issues, all Debian/Ubuntu. All on different nodes, all in different IP subnets. Going to try and look more into it in the morning to make sure we actually resolved and didn't just throw a bandaid at the issue.
Download speed from CacheFly: 51.4MB/sDownload speed from Coloat, Atlanta GA: 51.9MB/s
Download speed from Softlayer, Dallas, TX: 26.1MB/s
Download speed from Linode, Tokyo, JP: 11.1MB/s
Download speed from i3d.net, Rotterdam, NL: 6.65MB/s
Download speed from Leaseweb, Haarlem, NL: 48.6MB/s
Download speed from Softlayer, Singapore: 8.14MB/s
Download speed from Softlayer, Seattle, WA: 18.9MB/s
Download speed from Softlayer, San Jose, CA: 26.9MB/s
Download speed from Softlayer, Washington, DC: 57.6MB/s
everythings seems fine here, node jax01...
The fact the thread title and content has changed is disappointing.
@darknyan did you decide to butcher the thread, or did someone politely ask you to do so?
Eh, I butchered it. It was resolved and quite frankly I'm happy now. Although I was a bit annoyed I actually had to make a thread to get it resolved and I'm disappointed that more troubleshooting and steps to ensure this wasn't an isolated issur was not taken immediately, Skylar made a huge effort to get this resolved and I give him kudos for that.
Even so, you actually scored a big hit here because initially Crissic were saying you were wrong. To remove that is disappointing, and removes a trail for future buyers of how this provider had dealt with an issue.
Yeah, they fixed it in the end, but like you say it took this thread to do it, despite other customers apparently complaining about the same thing, which takes them off my consideration list personally.
If it bothers you and there still is a copy of what I originally posted somewhere, I'll revert it back, or perhaps the mods will.
@mpkossen @spirit If its still possible, could you please revert this thread back to the 2nd revision?
I definitely jumped to a conclusion too rapidly with this particular ticket, however without the thread we likely would not have known about the other users experiencing similar issues as we hadn't received any other tickets indicating a similar issue. As mentioned it impacted about 10 out of the many thousand containers we have so it wasn't something easily replicable or track able. Lesson learned with jumping to conclusions too rapidly however we did perform checks on the node and other containers on node/in same subnet and could not replicate at the time of the ticket.
http://webcache.googleusercontent.com/search?q=cache:439dGMZLFhMJ:lowendtalk.com/discussion/26751/crissic-network-speeds-not-up-to-par+&cd=7&hl=en&ct=clnk
Don't do owt for my sake, I'm just expressing an opinion. If you think it's the right thing to do then revert it. If not, don't.
It's up to you.
It's up to you.
Please
looks alright, but why to the west coast is too slow
Why don't you guys run upload speed tests as well?
Atrato is offline at the moment, which was running a large chunk of our network. Since atrato is having issues locally in JAX, our atrato uplink and our upstreams is off so we're maxing out some lines without some route tweaking to keep things in check until we can get our atrato uplinks back online. Should only be temporary (few days worst case). Some routes may be less than optimal until Atrato is brought back up and routing can be returned to their normal levels.
Original content restored per OP's request.
i also with crissic and the network seems fine. maybe there are blips in your VPS. maybe try to move to another node, if @SkylarM allow. It is just an alternative.
The thread. Did you bother to read it?
That 20ms ping looks weird. It's practically impossible for a server in the US to get a ping reply from Greece in 20ms.
@George_Fusioned Not really, it depends. here are some examples, some of them under 20ms, some higher. All servers in US:
All test from GRNET (Greek Research and Technology Network)
That is definitely not going to Greece
@jeffreywinters ? I think the script is a simple uploading and downloading script
@Jack Ok, I haven't search it. In your opinion, what is giving those results? -
@Jack I have about the same ping (around 200ms) from my home connection to CC and Crissic. So, what is giving wrong ping results to speedtest.net test?