Thanx for your suggestion. It is really value for me. On 2/18/07, Res <res@xxxxxxxxxx> wrote:
I think I might have read that wrong, dang, its sunday arvo afterall :) if thats 600K msgs a day on average, no sweat the below setup will do it as well. But it would help to know what software you will be running on it mbox or maildir, sendmail or qmail (with vpopmail? if not.. use it) On Sun, 18 Feb 2007, Res wrote: > Hi, > > On Sun, 18 Feb 2007, nilesh vaghela wrote: > >> I do not have idea, but I know we are goging to start with 3,00,000 user >> initially. >> >> let us say 20 per day per user so 20 x 3,00,000. Aproximate. > > Oh only that many messages? OK you could easily get away with something like > a HP DL380G4, use at least 2G ram, I'd use 2x36G 10K rpm scsi in raid 1 for > OS and server software dir, then 1x72G 10K rpm scsi for /var/log and > 1x36G 15K rpm scsi for /var/spool, and 2x300G 10K rpm scsi in raid 1 for > /home (assuming your mail store is under home) > If you use Qmail your /var/spool becomes /var/qmail. > If using MailScanner with spam assassin and anti virus, I'd suggest a ram > drive as well, use in this case 4G ram, with 2G locked for ramdrive. > > For 300K messages a day, the above will give you plenty of room to grow > as it will easily process far more than 300K. > > >> >> On 2/18/07, Res <res@xxxxxxxxxx> wrote: >>> >>> On Sun, 18 Feb 2007, nilesh vaghela wrote: >>> >>> > Any thing regarding server or servers. >>> > >>> > If any body can tell me the existing hardware setup for high end mail >>> > server. >>> >>> >>> >>> How much mail do you envisage processing per day? >>> >>> >>> > >>> > On 2/16/07, Res <res@xxxxxxxxxx> wrote: >>> >> >>> >> On Fri, 16 Feb 2007, Stephen Carville wrote: >>> >> >>> >> > >>> >> >>> http://www.oreillynet.com/pub/a/sysadmin/2007/01/05/fingerprinting-mail-servers.html >>> >> >>> >> Interesting article, but this doesnt represent much of the net as they >>> >> even admit, also certainly doesnt reflect the state of this country. >>> >> >>> >> Those figures are completely flawed, since theres a damn sight more >>> >> than 20K plesk boxes out there alone, not to mention all the other MTAs >>> >> running qmail without plesk and since qmail and vpopmail are superior >>> >> to anything in the hosted domains world, I can't beleive those figures >>> >> for qmail at all, thereby voiding the other results. >>> >> >>> >> The only way they can really do it, is to spend a month testing every >>> >> domain, results without dupe IP's showing close as unique as we >>> >> can really get. If the claim they can test 400K in a couple of hours no >>> >> reason they cant do the lot in a month. If its too hard, maybe they >>> can >>> >> ask netcraft. >>> >> >>> >> >>> >> >>> >> -- >>> >> Cheers >>> >> Res >>> >> >>> >> "We can be Heroes, just for one day" - Davey (Jones) Bowie >>> >> >>> >> >>> >> -- >>> >> redhat-list mailing list >>> >> unsubscribe mailto:redhat-list-request@xxxxxxxxxx ?subject=unsubscribe >>> >> https://www.redhat.com/mailman/listinfo/redhat-list >>> >> >>> > >>> > >>> > >>> > >>> >>> -- >>> Cheers >>> Res >>> >>> "We can be Heroes, just for one day" - Davey (Jones) Bowie >>> >>> >>> -- >>> redhat-list mailing list >>> unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe >>> https://www.redhat.com/mailman/listinfo/redhat-list >>> >> >> >> >> > > -- Cheers Res "We can be Heroes, just for one day" - Davey (Jones) Bowie -- redhat-list mailing list unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe https://www.redhat.com/mailman/listinfo/redhat-list
-- Nilesh Vaghela ElectroMech Redhat Channel Partner and Training Partner 74, Nalanda Complex, Satellite Rd, Ahmedabad 25, The Emperor, Fatehgunj, Baroda. www.electromech.info -- redhat-list mailing list unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe https://www.redhat.com/mailman/listinfo/redhat-list
