Why so many forum outages lately

[Greg]

OK There have been a number of issues over the last few weeks, alot of which arent really quickly fixable.

We are currently running an older version of MySQL than we would like due to a software bug with MySQL 5.0.20 and higher. It has been fixed in MySQL 5.0.23, however that has not yet been released.

The other issue is/was our monitoring of outages; this is currently being worked on and a new monitoring system is being deployed over the coming days that will enable more accurate monitoring of services.

So, essentially, we've had alot of problems with our Database system, mainly related to software bugs, which we have temporarily ironed out; however we are waiting for the 5.0.23 release of MySQL to completely resolve the MySQL crashing issues.

Also, when there has been an outage, it has taken us alot longer to resolve it due to a lack of a quality monitoring system. This has been addressed, and an SMS system has been implemented, and is in testing. It will go live before the end of the week.

We also had a DNS hiccup last night that caused both the site and forums to be unavailable (even though the servers were up). That has also been restored by the changeover to redundant DNS.

Yes, still issues with stability even though we are on our new killer platform - as for the upcoming new episodes, we should be all good for those. Our estimated restore time for full stability is mid June; so we should be fine way before the new episodes hit TV.

Peace out.

[Lord Zedd]

That is good news to hear Greg ! Hopefully soon 5.0.23 will be released soon.

[Pharaoh Atem]

thanks for the answer

[TechnoWraith]

Great news, Greg.

I take it our new hardware set-up is doing well for us, despite the software glitches?

[Greg]

We usually peak at 20% capacity, so yes there is ALOT of room to grow.

[CKO]

thx for that info Greg... lets just hope it stays up now ...

[Pharaoh Atem]

not sure of it's something you've been doing Greg but I've seen a huge improvement lately i.e. is much faster lately on the forums

[Greg]

Yes, we havent been playing around with things, thats why. There are usually *issues* with the servers when we are doing maintenance to them. We havent needed to touch any core/critical stuff over the last few weeks.

[Pharaoh Atem]

that's good hopefully there won't be anymore bumps in the road

[Pharaoh Atem]

OK There have been a number of issues over the last few weeks, alot of which arent really quickly fixable.

We are currently running an older version of MySQL than we would like due to a software bug with MySQL 5.0.20 and higher. It has been fixed in MySQL 5.0.23, however that has not yet been released.

The other issue is/was our monitoring of outages; this is currently being worked on and a new monitoring system is being deployed over the coming days that will enable more accurate monitoring of services.

So, essentially, we've had alot of problems with our Database system, mainly related to software bugs, which we have temporarily ironed out; however we are waiting for the 5.0.23 release of MySQL to completely resolve the MySQL crashing issues.

Also, when there has been an outage, it has taken us alot longer to resolve it due to a lack of a quality monitoring system. This has been addressed, and an SMS system has been implemented, and is in testing. It will go live before the end of the week.

We also had a DNS hiccup last night that caused both the site and forums to be unavailable (even though the servers were up). That has also been restored by the changeover to redundant DNS.

Yes, still issues with stability even though we are on our new killer platform - as for the upcoming new episodes, we should be all good for those. Our estimated restore time for full stability is mid June; so we should be fine way before the new episodes hit TV.

Peace out.

just wondering if any of the above bugs have been fixed yet things have been going smooth lately

[Greg]

The upgrades have been done, yes. We are now having issues with Apache - yay. It seems to be eating up RAM, then swap usage for no reason - or atleast, we havent figured it out yet.