$50 a year for a webiste that seldom
Originally Posted by BITSA,Jun 9 2006, 11:24 AM
Yet another say-nothing "statement."Memo to Dave and Mike: If indeed it is true that S2KI's technical difficulties are not ultimately grounded in a lack of money and therefore paid members (a possibility that I am more than just open to), then you need demonstrate to the readership the way or ways in which Cthree mismanages this site. Be (very) specific.
Originally Posted by zdave87,Jun 9 2006, 03:06 AM

That said, Dave, your "
" was an excellent way of saying that you cannot possibly demonstrate to the readership any sort of mismanagement on Cthree's behalf. Ergo, Dave, unless Mike is able to illustrate accordingly, then we can only assume that Cthree and xviper are indeed correct: that this forum's technical problems are clearly and firmly rooted in a lack of memberships and, therefore, money.https://www.s2ki.com/forums/index.ph...T&f=1&t=376100
***
Persevere.
Guy
Bumped for the edification of those able to show exactly how it is that S2KI.com is being "mismanaged" - those able to show that this site's technical problems are "not" grounded in a lack of individuals willing to pay for that which they use.
https://www.s2ki.com/forums/index.ph...T&f=1&t=376100
Thank you.
https://www.s2ki.com/forums/index.ph...T&f=1&t=376100
Thank you.
It looks to me that cthree is doing as much as he can to fix the problems.
cthree
cthree
[QUOTE]What changed is all of the servers were replaced, 4 new ones added, the httpd server changed from apache to lighttpd, php was changed from v4 to v5, mysql was changed from v4.1 to 5.0. Plenty of opportunity for lots of bitchy problems and lots of tuning issues like:- the new dual opteron database server won't run for more than 12 hours without hanging using both 32 and 64 bit versions of 2 different linux distributions. Hence why the site was going offline regularly. I switched it back to the dual Xeon it replaced which overloads it but at least it doesn't crash.
- lighttpd with 4 load balanced servers running php/fastcgi was running fine but was spitting out errors about clients timing out sending files. To eliminate the errors I increased the write timeout limit from 30s to 300s which worked but had the side effect of causing a large number of open and dead connections waiting for 5 minutes for clients that went away, that's why it was slow for the past couple of days. I made some tuning changes and fixed both issues and that's why it's fast again.
When I change parameters I don't know how things will change until I get a chance to see what happens over time so it can take a day or two to dial it in. the traffic at 1AM is not the same as at 1PM and what looks good in the evening could be big problems midday.
It's just the way it goes and yes it was most definitely broken.
cthree
It is my unfortunate #1 priority to have the site be as fast as possible. We are running on our old database server because I can't get the new one to not crash repeatedly. We'll reinstall yet another version of LINUX on it and see if it holds up. changing the backup server through off much of the optimization changes we made on other servers which performed horribly under the backup server. I had to re-optimize everything and I think I've finally found the sweet spot.
[QUOTE]What changed is all of the servers were replaced, 4 new ones added, the httpd server changed from apache to lighttpd, php was changed from v4 to v5, mysql was changed from v4.1 to 5.0. Plenty of opportunity for lots of bitchy problems and lots of tuning issues like:- the new dual opteron database server won't run for more than 12 hours without hanging using both 32 and 64 bit versions of 2 different linux distributions. Hence why the site was going offline regularly. I switched it back to the dual Xeon it replaced which overloads it but at least it doesn't crash.
- lighttpd with 4 load balanced servers running php/fastcgi was running fine but was spitting out errors about clients timing out sending files. To eliminate the errors I increased the write timeout limit from 30s to 300s which worked but had the side effect of causing a large number of open and dead connections waiting for 5 minutes for clients that went away, that's why it was slow for the past couple of days. I made some tuning changes and fixed both issues and that's why it's fast again.
When I change parameters I don't know how things will change until I get a chance to see what happens over time so it can take a day or two to dial it in. the traffic at 1AM is not the same as at 1PM and what looks good in the evening could be big problems midday.
It's just the way it goes and yes it was most definitely broken.
Originally Posted by zdave87,Jun 9 2006, 11:42 AM
It looks to me that cthree is doing as much as he can to fix the problems.
cthree
cthree
cthree
cthree
cthree
cthree
cthree
cthree








