I'm upgrading my server again and now that we have 8 disks we need to upgrade our power supply (I was told).
But do I really need a redundant power supply Or can I just buy ONE 650 watt to upgrade?
These are my current server specs:
2 X AMD opteron 270 dual core 2 X 2 GB kingston PC3200 DDR ram 2 X 1 GB kingston PC3200 DDR ram (6 GB total) 8 X seagate cheetah 15k.5 SCSI 73.4 GB, U320, SCA in raid 10 1 X LSI megaraid 320-2x raid controller, dual channel 1 X Tyan thunder 8KSE (s2892) mobo 1 X 2 * 4 port chenbro scsi backplane (8 ports total) 2 X AMD CPU cooler 2 X 40X40X10 fans (Currently 1 X zippy P2G-6510P, 2U, single power supply, 510 Watt) 1 X Sony FDD 1.44" 1 X chenbro RM311, 3U case with 8 hot swap bays 1 X LG DVD/CD-rom burner
do dual power supplies use more power than a single supply?
E.g. Say I have a server than uses two amps, powered by a single power supply. Now if I switch to a dual supply (and say each supply has the same efficiency rating as the single), does my server use more power? How much more?
My simple view of this is that it probably does, but maybe not much. The second power supply consumes some power itself, but since its not under load, it doesn't consume much. Therefore, my server with redundant supplies might use 2.1A or 2.2A.
I been invegistating into the 1U-3U Cases, and I been wondering about the power supply units that powers the boxes.
I know that Supermicro have very high efficiency power supply units in their chassis, but I am also wondering about the others, like Asus and Tyan chassis. Reason is that I am looking into purchasing servers, and I rather have those that have efficent power supply units in it than those stodgy Dell units, that is known for not quite efficient power supply units in it.
If anyone know of other 1U chassis that comes with efficient power units, I would like to know.
Just colocated at HE.NET and the the 15 amp limitation is a challenge. I'm just wondering what tricks there are to reduce power consumption by computer. Maybe people can add their tricks to this thread.
I'm experimenting at home right now. I got one of those KILL-A-WATT meters. (Egghead $18 - they are GREAT) And I'm measuring power usage with different configurations.
One thing I noticed is that the bios "cool and quiet" setting make a big difference. At idle current drops from 0.90 amps to 0.67. Of course if the processor is running 100% this probably goes away but if you have a cabinet of server and not all of them are maxed out all the time this should add up.
I'm using OpenVZ virtualization which allows my to use hardware more efficiently.
Just wondering if any of you have other tricks that cut electrical usage. Might have to put quad core processors in all my servers for expansion. Wish AMD would get their act together and get high speed quads out there.
the title says it all per month. just curious since this is something in europe that gets a price bump when it comes to power and space. yes i got two colo servers in different locations but never have i asked.
how often does a colo provider's datacenter go down? I'm not talking about resellers or their racks, but the primary provider itself.
This has been the 2nd time (this year I believe) that my datacenter at NAC has suffered a complete power outage [url], their backups failed, and my entire rack of servers were power-cycled.
Luckily I am not a web host but I am running some critical public web services/sites. I have all of the equipment to manage my own colocated machines from afar (monitoring, remote reboot hardware, and KVM/IP hardware for all of my machines) but I'm dead in the water if my datacenter's power is out.
I always ease my pain throughout a network outage or power outage by visiting DSLReports. Their HUGE website is hosted in the same datacenter (probably in the same room) as me and while it is a terrible thing to say, being able to share the downtime with a bigger fish is easier for me to handle.
Just picked up some used Dell PE 1750's, and was planning to add Drac's in them, but I can't seem to find which drac to use. Anyone that has any experience with this model?
Talked with a salesperson at Dell, first he didn't seem to know what Drac was, and after a while he came back to me with Drac IV which fits into PE1850. He belived it also should fit into 1750, but I don't belive that.
I have a couple of Dell 1950s and in one of them, I have 2x Seagate 15K.5s that I purchased through Dell and I also have a spare sitting in my rack in case one goes bad, also from Dell.
I was going to be repurposing one of my other 1950s and was going to get two more 15K.5s for it, but wasn't planning on getting them through Dell (rip off?). This way, could still keep the same spare drive around in case a drive went bad in that system as well.
When I was talking to my Dell rep recently when purchasing another system, their hardware tech said you can't use non-Dell drives with Dell drives in the same RAID array because of the different firmware between them.
Anyone know if it is true? Anyone have any experience with using drives from Dell in conjunction with the same model drives from a third party retailer?
I have been working on and designing a hosting company. Now setting up my names servers has came to a question. When I ran a previous business with a partner he only use Name Servers on 1 Machine in 1 location.
Now my question is my first server is going to be located in New Jersey, USA . So that is where My first set of DNS starts. With my previous partner I noticed most of the customers were based in Europe, Philippines, etc. Over seas. Now What I am asking is Will a VPS in a different location work for DNS Server. That way I can keep a lower budget till we get a few customers under out belt.
I currently run spiderloop.com a SEO control panel designed for windows hosting accounts. The site and web service is hosted on our own server and works great, however I am concerned about redundancy and do not want to purchase a whole other server just to be fail safe. So I am going to purchase a shared hosting account to facilitate redundancy.
So here is the question.
How do I do it? How do I set up my current host to redirect to the new host when it goes down?
If the current host has problems how can I force the users browser to redirect to the redundant host? I simply don't understand.
I've decided to move from managed hosting to colocated hosting and I'm in a bit over my head as far designing the server configuration goes. I'm looking for assistance with setting up a fully redundant configuration with no SPOF (single point of failure).
The colo provider ensures full redundancy to the cabinet, terminating with two network drops from diverse bandwidth providers and HSRP. I need to eliminate all SPOFs inside the cabinet.
1) I'm thinking of using the Cisco 2960 Series switch as the in-cabinet distribution switch. I believe it supports HSRP. From my understanding, the switch will automatically failover from one uplink to the other in case of a network problem on one of the uplinks, and this will all be transparent to any servers connected to the switch. Is this correct?
2) However, this introduces the switch as SPOF. So, I believe what I need to do is ask the colo for a second pair of network drops and deploy a pair of 2960's. Then I would connect each server to both switches via separate NICs. Is this correct?
3) Finally, assuming I am on the right track here, I would like this to all be transparent and automatic with no human intervention required to recover from a switch failure. The servers will be running RHEL 4. How would I configure the network interfaces on the servers so that they can transparently use either (or both) of the switches? Is there some way to assign the same IP address to more than one interface? How does the OS keep from getting confused about where to send the traffic?
Thanks in advance for any insight. I would also be happy to hear recommendations for excellent network consultants that I could hire to design and implement this configuration for me.
I have two servers. One hosts about 55 small sites and the other does nothing.
Can I make use of the spare one to take over automatically if the main server fails for whatever reason?
I had been thinking of using Rsync to backup accounts / mail every 6 hours or so and then manually changing the nameservers on the domains in the event of server1 failing. Would that even work? Then I got thinking about it all being automatic.
I have a few VPS's, the main one has cPanel/WHM and runs all my sites / email / DNS and MySQL DB's. Heres a little info:
VPS1 - CentOS 4.4, cPanel/WHM, runs all domains (OpenVZ) VPS2 - CentOS 4.4, Webmin, Slave DNS to VPS1 using Webmin cluster (OpenVZ) VPS3 - CentOS 4.4, Webmin, Slave DNS to VPS1 using Webmin cluster (Xen)
However, if VPS1 fails for say 24 hours, im screwed!
So, my question is can I get some kind of redunadancy built in somewhere. For example if someone is trying to access my domain "mydomain.com" and the main VPS is down, then the request for the site would go to VPS2, or VPS3...
The same for the mail server, if some is sending mail to one of the domains on the VPS, and the main VPS was down, the mail would be sent to my other VPS's.
I just don not like the fact there is a single point of failure!
I do have WHM managed Weekly and Monthly backups of all cPanel accounts etc.
Please give me the difference. Colo in carrier hotel, we can choose our preferred network provider, but should we do that if we cannot have our own tech in datacenter? How about the supporting service from carrier hotel? Just general question, cause I dont address exactly which facility.
And the second would be more expensive? Saying the same number of rack, amount of bandwidth... Who is providing IP addresses then?
We are a hosting company in Colombia, and we want start selling corporate email service.
The main idea is have two servers with redundance, I think this servers have to share the filesystem I don't know if using a scsi disk or just a network file system. We also want to have a good appliance for stop the SPAM, Im not so sure if we need to have redundance on it, its ideal but we are just starting the bussines and we dosn't have cliets yet.
Can anyone share their setup with us, and give us recomendations or tips?
Another option is, if someone know how to do that and have implemented this setup before and want to work $$ with us to implement that, please contact me to jbravo[at]colombiahosting.com.co
to setup a Windows server that can support the failure of the switch that comes before it in the network diagram. The idea would be that if the 15A circuit the switch was on failed, or the switch just died, the server would still be online.
I already have two switches in spanning tree just before the server. Is there a way to assign a single IP to the server on both its network cards and connect each card to a different (spanning tree'd) switch?
I'm in the process to setup a new service with an ISP with the following scenario and need your help.
I've got the rack (42U), servers and switches. Only routers has been left and here is that I need your help.
I have also 2 ports from the ISP where I can connect my routers. I need to get 2 router devices with auto sync feature in order to be able to setup a redundant plan in case that one of them goes down.
Those routers should have firewall features too in order to avoid setup iptables rules for each server. A basic DDoS protection is needed too.
I'm going to push around 100Mbit of traffic across the servers but that will happen after 3-4 months from the initial setup. In the first instance no more than 10-20Mbit will be used.
I heard a lot about Cisco but got no idea what model is the most suitable for my case. I will probably need a module for DDoS attacks and another one for advanced security IOS from what I read but it is not clear to me.
I'm looking to make high availability setup, and wondering how many of you have made it so? we are looking to multi-home the page with a round robin setup, using multiple VPSs/dedicated servers geographically different locations.
Right now i'm still looking at "stale" DNS setup, no automanagement of servers down. Is there a service/software which already offers automatic changes of zones for removing servers which are down, and adding them back when they get back online?
So the rule of thumb is if you have a 30A circuit to not exceed 80% or 24A.
If I buy a 30A circuit, I am assuming it is protected by a circuit breaker that is 30A, correct?
In *theory* we could run 29-30A all day and be fine.
So, I understand, the rule of 80% to be to the effect that "during startup, equipment uses more power ... or during runtime if it heats up, the fans go faster using more power" so it is safe to be at 24A in case you get spikes to 27-29A.
Is my reasoning correct?
Further, if you run primary/redundant circuits, is there really harm in running 13A on one circuit and 14A on the other.. and in the unlikely event of a failure, you are running 27A on one circuit (at 90% instead of 80%).
Any thoughts on this? Getting an extra 3A out of a 208V circuits is like putting in 3-4 more servers. If someone is doing this please let me know. I think it would work fine but i would like to know from the experts that do the colo stuff for a living what problems i could encounter, if any.