Anyone have any input on best practice for running vps servers it seems that there is a grey area when it comes to different configurations.
Ive seen so many
OpenVz
Hypervm with openVz
Virtuozo
Vmware
way to go about running a vps server and or servers and recomended configurations for each type.
Our colo has two carriers, call them A and B. I have discovered the colo provider is round-robining traffic out it's two carriers on a per-packet basis, not per flow.
Assume we want to reach destination IP a.b.c.d.
%> traceroute -q5 a.b.c.d
Results show that at the hop leaving the colo's border router, some packets transit Carrier A and some Carrier B, to the same destination IP, during the same traceroute.
Is this a routing Best Practice, or am I correct in thinking this is the Lazy Man's way of load balancing across multiple circuits, multiple carriers? BGP route selection does not seem to apply here (i.e., either Carrier A or Carrier B but not both at the same time).
Well I was quite surprised to see this but basically if you cancel a server at softlayer unless you go through after the fact and cancel all the associated orphans like ip ranges, vpn users, etc that was allocated to the servers you cancelled they will still bill you for this.
I have seen the notice on the cancellation process and I have always stated in cancellation ticket to cancel this server and all addons for it.
I find out today that I have been paying for at least 20 ip ranges, 15 vpn users and even a cpanel license for servers that I have cancelled months ago. How insane is that?
When a person cancels a server, they expect that server and all addons such as ip ranges, licenses and vpn users that was allocated to those servers to be cancelled as well. This seems not to be the case at softlayer and this is nothing more then an obvious way to bill clients for things that they are not even using or really cannot even use.
I guess that is what someone can expect from some wal-mart style network but I think this is simply an unethical practice. Most people would assume all items associated with a cancelled server would be cancelled with the server, that is what softlayer depends on .
IN the end I guess it is my fault for not reading "the fine print" but Im about sick of this fine print crap where hosts can outright decieve you and rip you off and then later point out some fine print validating their reasons.
Has anyone else had this issue at softlayer?
Do you access your server remotely (RDC) and administer your website or do you use SSL from any machine?
View 1 Replies View RelatedI am currently web hosting my own site (at home) through port 8000 using firewire. My activation is by invitation only. This means that I have to manually activate you before you can post on my board. I do this for security reasons and plus there is no need to invite the outside world. However, anyone on the outside world can view the site. The url is in the form of xyz.abc.uvw.rst:8000/forums/ where xyz.abc.uvw.rst is the IP address and 8000 of course is the port.
In order to make my site more secure, I have also changed the admin dir so that you can't just url to .../forums/admin/ and then start typing in passwords.
In addition, the version of my board is hidden and html is disabled.
I believe that I am secure but I would like to know other best web hosting practice methodologies that will allow me to be even more secure, particularly in the case where I am hosting on my own computer in my own home.
I have a ethics and legality scenario I would like to pose here. This is to be considered a fictional scenario and any similarity to any real company or person is unintentional.
COMPANY X has colo racks in DATACENTER and hosts a client named COMPANY Y. COMPANY Y also hosts 50 small web sites on cpanel and 40 VPS servers on their linux VPS node in exchange for a discount or other compensation. During the month of may and COMPANY Y's bill raises from being 3000 past due to over 12000 past due to various reasons including purchasing of hds from the DATACENTER using COMPANY X's account, tech work dont by DATACENTER for COMPANY Y needed also billed to COMPANY X's account, as well as BW and rack costs. Seeing this alrming bill by the end of May COMPANY X begins putting pressure on COMPANY Y to pay some invoices as no payments were received during the entire month of may and COMPANY X is also having a hard time paying DATACENTER due to not being paid by COMPANY Y but makes efforts to pay 25% of the outstanding bill in June.
Near the end of june COMPANY X has had enough and arrives at the datacetner to attempt to move theritr clients off COMPANY Ys hardware so they can suspend service of COMPANY Y completely. On the visit COMPANY X brings 2 new servers in and begins moving clients of their over and COMPANY X notices this action and begins to suspend and remove access to all of COMPANY X's clients. COMPANY X then decided to power off the boxes of COMPANY Y and removes drives containing COMPANY X customer data for offline recovery. and because they beleived the DATACENTER might be untrustworthy enough to release hardware fromt heir rakcs without their permission. The drives were for as much colatteral purposes as for recoverying customer data. They then pack the drives in boxes and a bag and leave the datacenter having spent nearly 20 hours working and head on a long trip home and signs out of the data center in front of various techs holding the server parts and tools.
Upon arriving back to their office/home
They were informed that the removal of the HDs was not allowed byt he datacenter because the account was past due and it was discovered that several drives with cOMPANY X's clients were left in the datacenter which COMPANY X now wishes they had remembered to get. DATACENTER is furious and demands the account be paid in full by a fixed time or the entire colo will be cut power. COMPANY X scrambles to find others to help them and another major client who has alot at risk decides to help and a payment of $7160 is made to DATACETNER but it came in a few minutes past the deadline. DATACETNER had suspended servuce, and cut power but stated they wil return service shortly. Shortly after this it is discovered that several servers belonging to COMANY X and two other companies who were supsended in the same day were not int he rack anymore. DATACENTER accout rep responds with the following message in ICQ/AIM .....
I'm looking for a method to basically clone my plesk site to another hosting provider also running plesk, the idea is so if my site A is down, I can just re-point the dns entry to Site B, and minimize downtime.
It makes sense to me to pay 2x hosting costs so I can have a site mirrored in case the main site goes offline, then I can just switch the dns to the new host. The trick is how can I automatically keep Site B mirrored as part of site A. Using Plesk Backup doesn't seem practical as it's not automatic. Does Cpanel have a site mirroring feature?
Would something like rsync work? Doesn't this require shell access? I'm looking for best practices to be able to mirror Site A to Site B in an automatic method..
Dedicated server has 2 HDD but I am not going to pay another $25/month for the hardware RAID solution (already stretched too far).
My plan is to install FreeBSD 6 and use Gmirror to establish a raid-1 "soft" mirror.
Advantages: Entire drive is mirrored including the OS. Drives can be remotely inserted or removed from the mirror set using a console command so its possible to uncouple the mirror and perform software updates on a single drive then re-establish the mirror only after the updates have proved successful.
Disadvantages: Lower I/O than hardware solution (not a problem for me) others???
I rarely see people consider software raid for a tight-budget server and I am wondering why? Could it be that other OS's dont have a solution as good as gmirror? Or is it just that crappy soft-raid in the past has left a bitter taste in admins mouths? Or perhaps admins need the extra I/O of hardware?