All CO.ZA problems to be resolved this weekend

Its always a wreck when asleep at the wheel. Good luck Uniforum.
 
They have done a pretty good job for the last 12 yrs or so that Ive been dealing with them. Considering the amount of data and the importance of the systems they run this is the first serious outage Im aware of and they have not lost any data which is a relief for all concerned! Good work Mark and Calvin and all the others in the back there.
 
Tommygun said:
They have done a pretty good job for the last 12 yrs or so that Ive been dealing with them. Considering the amount of data and the importance of the systems they run this is the first serious outage Im aware of and they have not lost any data which is a relief for all concerned! Good work Mark and Calvin and all the others in the back there.
Click to expand...

Any sysadmin would tell you a restore of systems that takes longer than 24 hours to do has a crappy or non-existent redundancy plan in place. It's just for 12 years they didn't have any mishaps (which is a brilliant track record) but it doesn't excuse the fact that they don't grow with the times or technology which inevitably leads to a disaster like this taking so long to fix in the first place. They're all to happy getting the R50 in for the couple of hundred thousand .co.za domains that exists, you'd think that their hardware would be hosted in a proper data center which wouldn't be affected by power spikes huh?

They probably trying to save money and hosting it in their offices...
 
O agree with AcidRazor. Surely 1 week downtime is indicative of a deeply flawed business continuity plan, especially considering the seriousness of their data records. Say for example they lost all their data records during this time...there'd be chaos (cos I'd be registering myadsl.co.za, mweb.co.za and other premium co.za domains long before anyone else :-) ).

So, Uniforum, get with the times. This past week should make you seriously consider your systems architecture and start taking redundancy and disaster management a bit more seriously. As the custodians of the co.za namespace, it is your d@mn duty to prevent this from happening again...no pressure :-)
 
Looking at it like that I guess you are right acid...however, even good recovery plans don't always work out despite spending lots of money on them. It would be interesting to know what systems they do have in place in terms of redundancy, fail over and redundant lines.
Any co.za ppl reading this thread?
 
Tommygun said:
Looking at it like that I guess you are right acid...however, even good recovery plans don't always work out despite spending lots of money on them. It would be interesting to know what systems they do have in place in terms of redundancy, fail over and redundant lines.
Any co.za ppl reading this thread?
Click to expand...

I doubt they waste their time reading. They probably only download porn/play online games during office hours, hence sounding rude when you phone in when you do have a problem, because you're interrupting their porn viewing or gaming schedule.

I agree that not all recovery plans ALWAYS work out. That is why you have redundancy in the first place. As an example, this is what I have:

Raid 10 configuration with 6 disks
Hourly backup of database/essential files
Daily backup of files that don't change as much
Daily incremental NAS backup
Weekly full NAS backup
4 week retention on hourly/daily backups
2 week retention of NAS backup
(btw, the NAS also configured in RAID and it ALSO gets backed up to a 3rd NAS system with a longer retention period)
Every week the NAS system gets backed up on tape
Every 2 weeks the NAS system's backup tapes goes to an offsite data storage facility

And the above is only for my web/database server. It's not even AS mission critical as their stuff.

What I want to know though is what data center are they using that is susceptible to power spikes????
 
AcidRaZor said:
What I want to know though is what data center are they using that is susceptible to power spikes????
Click to expand...

It appears Posix are the providers of not only the Uniform systems, but the hosting facilities as well. Not only that, Mark Elkins, the Technical Director of Uniform is also the Managing Director of Posix and WROTE the system over 12 years ago...

Posix has its own hosting facility in the Rosebank area strategically close to the main Internet Peering centre in Johannesburg. This allows Posix reasonably fast access to other ISP's in and around South Africa. There are seperate links to some of the other ISP's - eg SAIX (Telkom).
Our International Bandwidth also lands at this facility.

The Facility has backup generators and ech hosting cabinet is equiped with its own UPS - etc. The Hosting Centre is switched at 100Mb full Duplex. There are various security systems (Camera's, Keyed access - etc) to ensure the safety of your equipment. Machines are monitored and if your machine dies - you can be optionally (you decide and control) notified of the event. You can even temporarily remove your machine off the 'net if needed - all under your control.
Click to expand...
 
Tommygun said:
They have done a pretty good job for the last 12 yrs or so that Ive been dealing with them. Considering the amount of data and the importance of the systems they run this is the first serious outage Im aware of and they have not lost any data which is a relief for all concerned! Good work Mark and Calvin and all the others in the back there.
Click to expand...

+1 I guess I am more forgiving this time around because I had no changes or registrations in the last two weeks so I wasn't even aware they were having issues apart from the posts here. Hopefully they learn from this mishap.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X