RSS

News archive

Sporadic connectivity

Trouble at several stations in my area makes connectivity very sporadic. When there is connectivity it is only for a few minutes. My ISP is working on it

Back up

In case you have not noticed, we are back online. Things got a little more tricky than my ISP anticipated, resulting in no connectivity until early today (local time). Of course this coinsided with other things I had to attend to, which is why I could not write this 10 hours ago when things started getting back up.

Unscheduled downtime

During a routine reboot a data storage disk triggered a massive amount of unreadable file record segment reports. This made the reboot take forever an I eventually shut down the machine. Sorry for the inconvenience.

Outages

Appologies for the outages this week. Monday I started a new job where I work from home, and new equipment needed to be installed. Much to our surprise this gave alot of trouble as the new equipment needs to work together with existing equipment, and every time we took two steps forward getting things to work, we were put back two and a half. We are not fully done yet, but friday it looked like we finally got ahead. Hopefully the rest will be without further outages. As a side effect, when things actually run, it runs smoother and more secure than before.

Very slow stats update

Unfortunately the daily stats run for the 24th was extremely slow to finish - actually the first attempt never did, and I had to reverse everything and restart it. It was finally finished a couple of hours ago, at which time I set the run for the 25th in motion. This has the unfortunate side effect that the site has been hard to get through to all day yesterday and all day today - and it is not all over yet as it will soon be time for the run for the 26th. I have suspended the first incremental run for the day every day from now on, to prevent that from ruining suh slow daily runs. Slow daily runs is usually due to indexing of the database is getting too old - once the index has been updated the runs pick up speed again. Forcing such updates looks not to be an option available to me. Sorry for the inconvenience.

A small note on what is going on here

I know that somewhere on the site there are errors - I get the error messages, the messages just donot tell me where. I have ironed out a few with the help from users reporting errors, where they happen and what triggered them. When the system is running updates, especially the daily stats update, the old web-server is overburdened with requests, leaving you with 403-messages (too many users). When I notice this I restart the service as the queue will hang. Both of these issues should be cured on the new site, very slowly under development (everything on that site will be php on the new web-server). Both the new and old web-servers are running on a regular end-user PC - both will be moved to server-class hardware when all parts are bought and assembled. Stats-fans will have noticed a dramatic drop in team credits - this is due to a lost member at CPDN, and as we all know, CPDN removes and adds all user credits when they join/leave a team. Sorry to see a member leave. It looks like an up hill period for our team these days - but hang in there, we will get back into top 500 - and better. Update software is currently slowly being converted from php to windows native, speeding up the processes as bottlenecks will decrease and making use of multithreading where possible - this is also preparation for move to new hardware with more memory. Some may have noticed server status showing downloading xml most of the time recently. This is due to problems downloading from berkeley (seti and setibeta) - problems hitting not only esea.dk but other sites as well. Some projects have a few errors or inconistancies in their wap output, causing membership status to show up wrong here - it has been like this for a while (making our ministats members ranking wrong - not even I show up as a 100% member of our team). I will try to find a way to circumvent that. Ok, not that small a note. Anyway, thought you would like an update on these things.

Finally back

Sorry for the long wait, but the sites are now finally back online. The switch to a new computer was almost painless - unfortunately only almost and the thing that made it almost was inbound requests to the web server services. I have used all my spare time this week to find where the trouble was. Undoubtfully it had to be a setting somewhere that had changed without me saying so, but which one? Foolishly I did NOT look at the most obvious place until just now: Firewall settings. Now, in retrospect it is obvious that this had to be reset. During the move phase the OS couldn't detect the network adaptor on the new computer - and lost all connection settings for any adaptor that had been on the old computers. So, when using the adaptor I new would work (when finally finding the right driver) I naturally had to create a new connection for that adaptor. I did not realize that firewall settings were unique to each connection setting (or maybe I did not think that far as I have also been ill this week). Anyway: We are back :)

Todays maintainance

As mentioned earlier this week, I planned to replace the PSU on the computer this weekend, and I had planned to do some other maintainance while the computer was down anyway. Neither of those was done, as it turned out I had trouble getting it running again. Time was spent finding the source for this. Apparently the video board is getting worn out. After much trouble getting the video board to comply and getting the computer running, I decided that this is as far the road goes for this computer. The amount of things not going well, in need of replacement etc. is just mounting up, so I will let it do its final work until next weekend. I had already planned to replace this computer anyway and add a real server to take care of web, updates etc. The server is still a bit in the future, though, and the replacement workstation was supposed to be a self build one, but things sometimes do not go as planned. So, saturday I will get a new computer, and the system will be taken down while moving hard drives. Hopefully that will go much smoother than those things usually does.

Downed computer

Sorry for the inconvenience, but the computer was off when I came home - do not know for how long (if anyone knows, please inform me). Not only was the computer off, but it would not start again, until I had it physically moved to another power source - moving it back and it turned on. I am guessing that the power supply is about to go, so I have turned off intensive software that produces heat (like boinc). Replacement of the PSU will happen this weekend, if I can stretch the life of the current one that far.

boinc.exe crashes, update

The cause of the problem is found. If the size of an accoun_*.xml file exeeds 64kb, boinc.exe will crash. For many this file for QCN has done that. As far as I am told boinc-developers are working on a fix (wether that fix will be to prevent the crashes or to accept larger account files, I do not know), and QCN has already made a fix to prevent that the file exeeds the 64kb boundary. Meanwhile here is two fixes you can do to get boinc.exe running again: 1) Delete the QCN account file - this does detach your client from QCN 2) Edit the QCN account file, simply delete from the first quake-tag to the last quake end-tag (including those tags) - this will not detach your client from QCN.

boinc.exe crashes

If some of you are having trouble making boinc.exe keep running, I can inform you that the matter is being looked into - and you are by far not alone. More on this when I know more.

Recent trouble

A while ago I got tired of the mouse connected to the computer, as it was not working properly any more and it was only getting worse. So I bought a new one. It did not take many hours after I connected the new one before I experienced a spontanious reboot. Of course I emidiately suspected the new mouse, but a collegue of mine convinced me that it had to be due to heat. Since then I have been trying to keep the computer cool which has been hard most of the time. As time past, I got more and more convinced that it had to be the new mouse - the computer still did spontanious reboots, and it was always when I was home and working on the computer, and always when I clicked the left mouse button. Yesterday I got completely sure - the coolest day so far this summer, absolutely no heat problems at all, and yet again the computer rebooted when I clicked the left mouse button. I changed to another mouse and since there has been none of that. However I needed to shut down briefly this evening, just a simple reboot to complete an installation, and had trouble getting things running as they should again, so that simple short reboot turned into hours of downtime. Got it fixed. However, expect downtime saturday, as I will do some cleaning of the interior of the computer.

Downtime next weekend

Next weekend I will add some power surge protection, which means sites and services will be down until things are ready. Better grounding and double protection. I will keep downtime as brief as possible. On another note, mentioning down time, you probably noticed that things still go unreachable around here from time to time. Clearly the solution I made does not work. Problem is I am not that experienced yet programming for Windows, so please bare with me. I will see if I can fix it tonight, if not, I will create a temporary alternate solution, which will simply restart apache once every hour no matter what.

PSU replacement postponed

Sorry about the delayed message, but the replacement of the power supply announced earlier has been postponed to this upcoming saturday.

Downtime

This weekend I will replace a PSU which means downtime for approx. an hour at the most. I expect to do this tomorrow Saturday, but it may be postponed to Sunday.

Disconnected?

It would seem so, but no, I do not get disconnected time and time again. I do not know what is going on, but I get the same problem with apache as with IIS - web server service goes balistic, running either 50% or 100% on the CPU. Maybe I am subject of DoS attacks, I am not sure. Either way, I am sorry for the inconvenience this may cause you, but I am working on a temporary solution that restarts the service when the process involved is using too much CPU at a prolonged period of time. Should be in place some time this weekend at the latest.

ISP changes at april 28th

As my current ISP has been purchased by a competitor, I have chosen to formaly change ISP now, to the ISP that has bought my current one. This is due to a better offer (price vs. speed vs. service) and a better control over the transition. This means change of equipment and IP which will happen april 28th. So expect downtime at that date. I will naturally try to keep the downtime to a minimum - and note that the IP changes as well, which means it could take a while for that to propagate.

Facebook group

ESEA now has a facebook group. If you are on facebook, please join us there.

Changes to server

I have made some changes that may affect you when trying to reach some pages directly. All ASP-based pages are now on port 8080. So to reach those pages you need to add :8080 to http://www.esea.dk (as in http://www.esea.dk:8080). It is easier just to go http://www.esea.dk/esea and brows to the wanted page. Later the ASP pages will be converted to PHP which can be reached normally.

Irregular updates

All updates of every kind are disabled at my server at the moment, and are only performed when I manually start them. This is due to some problems with updates qeueing up and slowing everything down, making it impossible to work. I am currently trying to uptimize some parts and de-uptimize others (de-uptimizing xml-downloads so they don't take up all bandwidth and qeueing up other outgoing updates)

RSS

News archive