ViRGE
Elite Member, Moderator Emeritus
Guys, we need you to make sure that your Dnet clients are both up to date, and configured correctly.
For those of you using the TA pproxy, we've had to implement an auto-blacklist because of clients repeatedly hitting the machine, dragging it down in terms of connections. Right now, I'm looking at 2 dozen or so IP addresses that are banned for an hour because their clients kept hitting the Core at a ridiculous rate. Please check all your clients and make sure they're at least version 468, if not later than that, as the clients before that are a large source of the hits. If you're curious about what machines are ending up on this list, we'll have more information once we get email addresses attached to the IPs in question.
Also of note: please check your OGR stub count settings. We had a user with a pproxy download 5000 stubs last night, which kept us drained through the night(as getting stubs from Dnet is just as difficult for the same reasons). OGR stubs take far longer to process than even the largest RC5 WU's; if you want a number to work with, start with 10 stubs, and work from there. Those 5000 stubs will take a moderately sized herd months to process, and might go stale before that. Please only take the OGR stubs you need for now until we can get a steady flow of stubs.
Lastly, some good news. With RC5 over, we're tweeking the stat times on the TA Website so that OGR runs are more frequent. Right now, we're looking at having runs every 10 minutes, since OGR logs process so much faster. If you're not already using the TA PProxy, now is a good time to start; proxy.teamanandtech.com is open for buisness, accpeting connections on the normal port(2064) along with port 23. Stats are availible at http://www.teamanandtech.com/dnet/.
Thank you for your attention and support(and fixing your clients), and good luck to the TA OGR team.🙂
For those of you using the TA pproxy, we've had to implement an auto-blacklist because of clients repeatedly hitting the machine, dragging it down in terms of connections. Right now, I'm looking at 2 dozen or so IP addresses that are banned for an hour because their clients kept hitting the Core at a ridiculous rate. Please check all your clients and make sure they're at least version 468, if not later than that, as the clients before that are a large source of the hits. If you're curious about what machines are ending up on this list, we'll have more information once we get email addresses attached to the IPs in question.
Also of note: please check your OGR stub count settings. We had a user with a pproxy download 5000 stubs last night, which kept us drained through the night(as getting stubs from Dnet is just as difficult for the same reasons). OGR stubs take far longer to process than even the largest RC5 WU's; if you want a number to work with, start with 10 stubs, and work from there. Those 5000 stubs will take a moderately sized herd months to process, and might go stale before that. Please only take the OGR stubs you need for now until we can get a steady flow of stubs.
Lastly, some good news. With RC5 over, we're tweeking the stat times on the TA Website so that OGR runs are more frequent. Right now, we're looking at having runs every 10 minutes, since OGR logs process so much faster. If you're not already using the TA PProxy, now is a good time to start; proxy.teamanandtech.com is open for buisness, accpeting connections on the normal port(2064) along with port 23. Stats are availible at http://www.teamanandtech.com/dnet/.
Thank you for your attention and support(and fixing your clients), and good luck to the TA OGR team.🙂
