Thread: Replacement for developer.postgresql cvsweb
After continued frustration with the performance of cvsweb at developer.postgresql and what appeared to be prolonged downtime yesterday -- I was getting timeouts -- I spoke with Omar about setting it up on another machine. You can access it at: http://treehou.se/cgi-bin/cvsweb.cgi/ It seems to be much faster. This machine is hosted in the US. More horsepower can be thrown at the problem if need be. Ideally, we'd be able to set up a domain such as cvsweb.postgresql.org and point it at treehou.se and then modify the committers emails. At the moment, treehou.se is pulling cvs (with cvsup) once an hour. I think we should pull more regularly, but I'm not sure just how regularly. Ideas? Thanks, Gavin
"Marc G. Fournier" <scrappy@postgresql.org> writes:
> On Sun, 21 Aug 2005, Tom Lane wrote:
>> Marc, why is it that cvsweb has gotten so remarkably slow lately? I
>> agree with Gavin that the performance is pretty unacceptable --- I just
>> don't agree that moving it offsite is the answer.
> Is it still? Things should be okay as of this aft, since I moved the VM
> to mars ... now that things are 'settled down' from this morning, I'll go
> over mars and make sure things are 'up to speed' over there as well ...
I think you missed the context --- this is not about last night's
outage. cvsweb has been pretty sucky for weeks :-(
regards, tom lane
Gavin Sherry <swm@alcove.com.au> writes:
> Ideally, we'd be able to set up a domain such as cvsweb.postgresql.org and
> point it at treehou.se and then modify the committers emails.
Personally, I'd be pretty dang unhappy with a cvsweb facility that
reflects hour-or-so-behind-the-times data. And even more so with
changing things so that you couldn't immediately click on commit
messages to see what had been done.
> At the moment, treehou.se is pulling cvs (with cvsup) once an hour.
> think we should pull more regularly, but I'm not sure just how
You do know that the anoncvs mirror only updates once an hour...
Marc, why is it that cvsweb has gotten so remarkably slow lately?
I agree with Gavin that the performance is pretty unacceptable ---
I just don't agree that moving it offsite is the answer.
regards, tom lane
On Sun, 21 Aug 2005, Tom Lane wrote: > Marc, why is it that cvsweb has gotten so remarkably slow lately? I > agree with Gavin that the performance is pretty unacceptable --- I just > don't agree that moving it offsite is the answer. Is it still? Things should be okay as of this aft, since I moved the VM to mars ... now that things are 'settled down' from this morning, I'll go over mars and make sure things are 'up to speed' over there as well ... To Gavin ... the 'prolonged downtime' this morning was the result of jupiter losing two drives , and it going down ... I have some spare drives here that I'll be shipping down this week, to fix it ... as soon as we knew (within an hour), we brought the VM up on its backup server, and today we moved it off of that server and onto a less loaded one, until I can get those drives down and jupiter rebuilt ... ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Sun, 21 Aug 2005, Tom Lane wrote: > "Marc G. Fournier" <scrappy@postgresql.org> writes: >> On Sun, 21 Aug 2005, Tom Lane wrote: >>> Marc, why is it that cvsweb has gotten so remarkably slow lately? I >>> agree with Gavin that the performance is pretty unacceptable --- I just >>> don't agree that moving it offsite is the answer. > >> Is it still? Things should be okay as of this aft, since I moved the VM >> to mars ... now that things are 'settled down' from this morning, I'll go >> over mars and make sure things are 'up to speed' over there as well ... > > I think you missed the context --- this is not about last night's > outage. cvsweb has been pretty sucky for weeks :-( Ya, that is something that I was working on, in fact, last night before last night's outage, and something I have a new server here in the office to finish putting together to deal with more long term on ... its taken a bit longer to get this new server down then originally planned, sorry :( ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664