On 9/19/07, jon robles <[EMAIL PROTECTED]> wrote:
> Hello PLUG,
>
> Can you recommend machine specs for a CentOS running postfix that processes
> around 5-8M mails/day? Thanks
>
> Jon
> _
> Philippine Linux Users' Group (PLUG) Mailing List
> plug@list
Yes, mail traffic is almost impossible to predict, so we are planning to
deploy servers that can handle twice the size of the average traffic. But
our dilemma is it's virtually impossible the growth of spam, so we might as
well hope for the best and plan for the worst.
We actually separated our RB
The bad part is, once you get it working, "barely," everytime someone
does a DDoS or spam run on you, you will get massive queues AGAIN.
On 9/21/07, Ian Dexter R. Marquez <[EMAIL PROTECTED]> wrote:
> Orly's right. A "telco-grade" SMTP infra is the way to go. Or go for
> the mail security applianc
minute is like an equivalent of 40 seconds.
>
> Has anybody have an idea? Is this a possible hardware issue?
It is probably not a hardware issue. My notebook is afflicted with
the same problem. However, I have not noticed any similar problems
when running WinXP. The issue could be related to AC
On 9/19/07, thad <[EMAIL PROTECTED]> wrote:
> I have the same type of mobile drive though its Maxtor. I just added
> NTFS support and works fine with Fedora 7 and it got detected. Last
> weekend I tried mounting a 300GB SATA with generic made in china
> bought from ebay. To test it I initially I pl
On 9/20/07, thad <[EMAIL PROTECTED]> wrote:
> On 9/20/07, andrelst <[EMAIL PROTECTED]> wrote:
> > Have seen this problem in vmware virtual machines where linux is a
> > guest. The fix is to add a parameter in grub/lilo. Check google if
> > there is a fix.
>
> The box is not set up to host any virtu
On Friday 21 September 2007 05:15:19 thad wrote:
> Has anyone encountered a system clock that run too fast? The machine
> is a ProLiant DL580 G3 4- way dual core running on RHEL4 64 bit. The
> ntp is configured with local time server and it is peering. But the
> issue it is always getting ahead of
On 9/20/07, andrelst <[EMAIL PROTECTED]> wrote:
> Have seen this problem in vmware virtual machines where linux is a
> guest. The fix is to add a parameter in grub/lilo. Check google if
> there is a fix.
The box is not set up to host any virtual machine, its a database server.
>
> NTP won't help
does the clock really sync during synchronization or not all?
if the former is the case, then you need to force sync the
machine by using ntpdate. otherwise, it would keep on doing
a step sync.
On 9/20/07, thad <[EMAIL PROTECTED]> wrote:
>
> Hi,
>
> Has anyone encountered a system clock that run
Have seen this problem in vmware virtual machines where linux is a
guest. The fix is to add a parameter in grub/lilo. Check google if
there is a fix.
NTP won't help you here because it tries to change the time gradually,
but then the time drifts again and so on and so forth. Besides, NTP is
smart
Hi,
Has anyone encountered a system clock that run too fast? The machine
is a ProLiant DL580 G3 4- way dual core running on RHEL4 64 bit. The
ntp is configured with local time server and it is peering. But the
issue it is always getting ahead of time that we observed that 1
minute is like an equiv
FTA:
http://www.networkworld.com/news/2007/092007-open-source-unavoidable.html
"To some of you, this is shocking, you don't believe this,"
"Others are saying, 'Forget 2011, it's already here today.'"
.
"Open source is going to come into your network whether you like it or not."
Orly's right. A "telco-grade" SMTP infra is the way to go. Or go for
the mail security appliances.
I recall having a customer with nearly that amount of mail traffic.
They had a problem reining in the queue, even with their big iron
boxes. Couple that with having AV software acting as the mail pro
On 9/20/07, Michael Tinsay <[EMAIL PROTECTED]> wrote:
>
> mdadm did not find any error with the disk, so I did mdadm --re-add. Took
> an hour to sync/update the disk, but it seems all is ok now.
>
Cool. Great that you recovered from that. It could be a hellish
experience, specially for mission-cr
guys,
tnx po sa reply ill check on it and give you feedback tnx
- Original Message
From: jan gestre <[EMAIL PROTECTED]>
To: Philippine Linux Users' Group (PLUG) Technical Discussion List
Sent: Thursday, September 20, 2007 6:26:55 AM
Subject: Re: [plug] squirrelmail problem
Chec
I have a friend over here at the province who was able to configure
his globelines broad band modem as bridge. The router that is doing
pppoa is a dlink. The modem? I didn't bother to ask.
I don't think you can do both pppoa and pppoe. If your area uses
pppoe, then you cant use pppoa, vice versa.
Bridged mode on prolink 9000 will never work with globelines. and
there is no firmware update available.
anybody who wants to stay sane is better off with a speedtouch
connected via usb to a linux box via pppoatm.
the prolink modem is crap. its so unstable its driving all of us up
the wall. when
What you are referring to is wear-levelling. If the write-erase cycle is
happening frequently on a group of sectors while other sectors remain
slightly used then you have uneven level of wear. In that case you dont
need a journaling fs but a wear levelling fs that will dististribute the
wear e
I have a problem connecting to globe dsl using our prolink h9000b modem.
when i set it to "1483 bridged ip llc" rp-pppoe times out waiting for
a "pado" packet. I can connect to the internet with the the built in
pppoa vc-mux.
does that means that i need to use pppoatm in my linux router?
--
Lay
Email is a highly disk-bound activity.
8M emails per day including spam- and virus-filtering is no joke.
Back at mozcom before we were only doing 1M emails per day and we had
like 8 machines for that (albeit single-processor Athlons).
Personally I think the traditional SMTP infrastructure like P
but... my understanding of the necessity of journaling on flash is
that, if you use MSDOS FAT filesystem, the FAT sits on a particular
spot on the device. So when you add files, delete files, etc. that
portion of the device gets written and over-written all the time.
Which would make it fail that m
actually the postfix are just for relaying mails to two separate systems. It
just actually serves as relay with rbl enabled using spamhaus and spamcop. I
already have 5 mchines in place but a traffic of 5.5M (with rbl enabled) or
more causes hellish queues.
Here are the general specs of the machin
On 9/20/07, jon robles <[EMAIL PROTECTED]> wrote:
>
> Do you have any idea on how many servers should we have for this kind of
> traffic?
i'm no expert but i think you should set up a clustered mail server, a
minimum of 3 server will suffice, i think? the first server just handling
smtp while th
23 matches
Mail list logo
