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
... Or tending to a tribe of children and playing schoolmaster.
@Voss I started my art exam a couple days ago(time consuming), and had deliveries to check at the DC, managed clients to phone.
I'm pretty busy at the moment, but managing fine.
Just had a outgoing attack from one of our coventry servers, we nulled the IP and cleared arp cache on the switch.
We'll be getting eth0 and eth1 bonded on the server your on @cause, should prevent someone consuming the majority of the port.
Lol, you're still at GCSE right? Shouldn't you be worrying about your proper subjects? You must have proper exams to be revising for this May/June.
I have done my maths, majority of my GCSES are coursework based which I did 1-2 Months ago.
RE i failed, but i'm retaking it.
Art and French I have C's definitely.
I can also leave in May and come back for my exams.
Don't kids have to stay in school till there 18 now? I'm glad i missed that train.
What, did you drop out lol?
I'm 22 lol, this only came in a few years back.
@Jack, thank you for the clarification.
https://www.dropbox.com/s/qo4zndnctg1oa89/ping_packetloss_overview-pinpoint=1360345717,1360366507.png
and received alerts at 4:48, 5.06, 6:06, 6:27 in JST(+0900). his comment (/discussion/comment/204308#Comment_204308) was posted at 5:46AM in JST.
Could anyone let me know why bonding prevent consuming bandwidth?
Well unsure of the offer details but bonding it to aggregate bandwidth. (i.e. 2 100Mbit = 200Mbit total).
I'd think these are 100 speed links from the server to the switch, but could be a busy node with saturated gigabit.
Bonding is to make more bandwidth available, more throughput.
@pubcrawler is correct.
Thanks.
I just thought if someone can consume whole port(s) on the server, 2 port aggregation between server and switch would simply double attacker's traffic and leave the attacker as monopolist.
Only if they had the bandwidth to saturate the bond
That's exactly what would happen if the VM is not limited to 100mbps itself.
Hard to say why bonding is occurring.
I use it for redundancy and for storage here. Never can have fast enough.
Odds are they are running 100 port speed. Overuse on Gbps happens though if high enough customer density and actual use.
We had ppl going at 800-1000 flat without DDoS. There can be throttling and limiting, but better watch the nodes and allow bursts from time to time, normal load on 1 gbps port is some 100 mbps, there is place to grow, limiting the machines might be a solution when you cannot watch all the time, tho.
@Jack I think thats what was causing the high loads on a node earlier, sorted anyhow.
Yesterday, in my timezone, got some service down alerts. Gmail joined 17 messages into one thread.
https://www.dropbox.com/s/rl3j81c0en25qe3/ping_packetloss_overview-day.png
@Jacob said these outage were caused by hardware failure and outbound DoSes. I wonder what could be covered by his SLA.
And also he said I have to answer something important to solve the problem, in private, but still I dont get what that was.
In my eyes, @Jacob was trying to hide anything unwanted for him.
@cause your just trying to get something for free here!
@cause How are things 4, nearly 5 days later? I'll also update this later in the month, and check up how things are.
My monitoring servers said last network issues were on Feb 11 . These outages were not so long but 3 service down alerts were sent. after that, i got only one alert yesterday.
IMHO, post your own monitoring log would be better than ask someone if you do not believe one's words.
According to a notification mail from them on Feb 11, they did not throttle burstable port speed so anyone could consume whole bandwidth of the hostnode. now it seems limited to 100Mbps per container.
I believe what you're saying it's just our monitoring being a PITA for now, untill I have time to work on this which won't be for the next few days you will just have to set them to spam or something.
I am looking at it now, and will have a short play with it and see what I can do.
Other then that, I'm referring to the actual service you are recieving (disregarding the service alerts), how is the uptime, and the overall performance of your VPS.
I don't want this thread to end negatively, but it may still be to early to judge.
Also at the slightest hint of packetloss the monitor will flip, I have adjusted some values on it but not sure how this will hold out, it does triple check downed hosts but again it's PITA at times.
There was a few containers with unthrottled ports, and they was burning through bandwidth.
Feb 10Th - Uptime 99.65% , Downtime 5 Minute(s)
Feb 11Th - Uptime 98.96% , Downtime 15 Minute(s)
After that it is all clear, and no problems!