I could do with some suggestions on web hosting, if this is not an 
appropriate place to ask please move me on!

I want to host a site that will have lots(hundreds, perhaps eventually a 
few thousand) pictures - jpeg format 300x300 pixels. They will be good 
quality and need to display as such. It could be updated several times a 
day 6/7 days a week.

My own site is with Heart Internet and I have noticed recently they are 
beginning to creak a bit and suffer from breakdowns/down-time, I know I 
can't get 100% but I need as near as possible.

Any thoughts? I can' say money is *no* object but it is secondary to 
reliability and speed.

Many thanks.

-- 
Jeff Gaines Damerham Hampshire UK
640k ought to be enough for anyone.
(Bill Gates, 1981)

Jeff Gaines wrote:

> I could do with some suggestions on web hosting, if this is not an
> appropriate place to ask please move me on!
>
> I want to host a site that will have lots(hundreds, perhaps eventually a
> few thousand) pictures - jpeg format 300x300 pixels. They will be good
> quality and need to display as such. It could be updated several times a
> day 6/7 days a week.
>
> My own site is with Heart Internet and I have noticed recently they are
> beginning to creak a bit and suffer from breakdowns/down-time,

Loss of service you mean ?


> I know I can't get 100% but I need as near as possible.
>
> Any thoughts? I can' say money is *no* object but it is secondary to
> reliability and speed.

How much do you pay now ? My brother in law does hosting if you want a
quote. He's also looking at a limited back-up server to maintain vital
aspects like email and a basic web presence should the main server go down
(as they do from time to time).

Graham

In article ,
Eeyore   wrote:
>
>
>Jeff Gaines wrote:
>
>> I could do with some suggestions on web hosting, if this is not an
>> appropriate place to ask please move me on!
>>
>> I want to host a site that will have lots(hundreds, perhaps eventually a
>> few thousand) pictures - jpeg format 300x300 pixels. They will be good
>> quality and need to display as such. It could be updated several times a
>> day 6/7 days a week.
>>
>> My own site is with Heart Internet and I have noticed recently they are
>> beginning to creak a bit and suffer from breakdowns/down-time,
>
>Loss of service you mean ?
>
>
>> I know I can't get 100% but I need as near as possible.
>>
>> Any thoughts? I can' say money is *no* object but it is secondary to
>> reliability and speed.
>
>How much do you pay now ? My brother in law does hosting if you want a
>quote. He's also looking at a limited back-up server to maintain vital
>aspects like email and a basic web presence should the main server go down
>(as they do from time to time).

One of my servers:

 18:51:06 up 1075 days,  4:53,  1 user,  load average: 0.16, 0.03, 0.01

Just 20 days until its 3-year birthday :)

Gordon

On 28/09/2008 18:50, Gordon Henderson wrote:

> One of my servers:
> 
>  18:51:06 up 1075 days,  4:53,  1 user,  load average: 0.16, 0.03, 0.01
> 
> Just 20 days until its 3-year birthday :)

Are you going to let it have lots of patches as a present?

In article <HdGdnSdDqdRuWULVnZ2dnUVZ8u2dnZ2d@posted.plusnet>,
Andy Burns   wrote:
>On 28/09/2008 18:50, Gordon Henderson wrote:
>
>> One of my servers:
>> 
>>  18:51:06 up 1075 days,  4:53,  1 user,  load average: 0.16, 0.03, 0.01
>> 
>> Just 20 days until its 3-year birthday :)
>
>Are you going to let it have lots of patches as a present?

Why wait until then?

  # apt-get update
  ... etc ... including
  Hit http://security.debian.org sarge/updates/main Packages
  ... 
  Fetched 346B in 2s (154B/s)
  Reading Package Lists... Done
  # apt-get upgrade
  Reading Package Lists... Done
  Building Dependency Tree... Done
  0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

It's fine. It's running a few fairly specific applications and low-volume
websites, it's firewalled and not really that general-purpose, but other
than that, it's sitting conencted to the Internet at large and humming
quietely to itself.

Debian Sarge is still being maintained, so I'm happy. I do still have
a server on Debian Woody which had over 3 years uptime until just a few
weeks ago when the motherboard Ethernet failed for no apparent reason )-:
However it's not exactly doing anything critical these days.

Gordon

On 28/09/2008 19:43, Gordon Henderson wrote:

> Why wait until then?
> 
>   # apt-get update
>   0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

I only run one Debian box, had to upgrade to Etch to use a newer version 
of iptables, it has been very hassle free, but still surprised Sarge 
hasn't needed any kernel update in 3 years.

In article <fuydnfn2NItmTkLVnZ2dnUVZ8srinZ2d@posted.plusnet>,
Andy Burns   wrote:
>On 28/09/2008 19:43, Gordon Henderson wrote:
>
>> Why wait until then?
>> 
>>   # apt-get update
>>   0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
>
>I only run one Debian box, had to upgrade to Etch to use a newer version 
>of iptables, it has been very hassle free, but still surprised Sarge 
>hasn't needed any kernel update in 3 years.

It's not the default sarge kernel, but one I deemed stable enough at the
time to compile for the hardware. (Dell) Who can actually login and get
a shell is half the battle and that's very tightly controlled here.

Gordon

Jeff Gaines wrote:
> 
> I could do with some suggestions on web hosting, if this is not an 
> appropriate place to ask please move me on!
> 
> I want to host a site that will have lots(hundreds, perhaps eventually a 
> few thousand) pictures - jpeg format 300x300 pixels. They will be good 
> quality and need to display as such. It could be updated several times a 
> day 6/7 days a week.
> 
> My own site is with Heart Internet and I have noticed recently they are 
> beginning to creak a bit and suffer from breakdowns/down-time, I know I 
> can't get 100% but I need as near as possible.
> 
> Any thoughts? I can' say money is *no* object but it is secondary to 
> reliability and speed.
> 
> Many thanks.
> 
Have a look at claranet: they are fairly professional.

Gordon Henderson wrote:
> Debian Sarge is still being maintained, so I'm happy.

Not by the Debian Project:

<http://lists.debian.org/debian-announce/2008/msg00001.html>

In article <gbots9$1ckm$1@energise.enta.net>,
Paul Cupis   wrote:
>Gordon Henderson wrote:
>> Debian Sarge is still being maintained, so I'm happy.
>
>Not by the Debian Project:
>
><http://lists.debian.org/debian-announce/2008/msg00001.html>

Hm. Must have missed that one. Oh well, c'est la vie.

The servers still not broke, so I'm still not going to fix it, if that's
what your point is.

Gordon

Gordon Henderson wrote:
> In article <gbots9$1ckm$1@energise.enta.net>,
> Paul Cupis   wrote:
>> Gordon Henderson wrote:
>>> Debian Sarge is still being maintained, so I'm happy.
>> Not by the Debian Project:
>>
>> <http://lists.debian.org/debian-announce/2008/msg00001.html>
> 
> Hm. Must have missed that one. Oh well, c'est la vie.
> 
> The servers still not broke, so I'm still not going to fix it, if that's
> what your point is.
> 
> Gordon
Indeed. Dont. However in a fit of madness  to get my dev machine running 
what the real machine was I upgraded to etch..I have to say it just 
...worked!

In article ,
The Natural Philosopher  <a@b.c> wrote:
>Gordon Henderson wrote:
>> In article <gbots9$1ckm$1@energise.enta.net>,
>> Paul Cupis   wrote:
>>> Gordon Henderson wrote:
>>>> Debian Sarge is still being maintained, so I'm happy.
>>> Not by the Debian Project:
>>>
>>> <http://lists.debian.org/debian-announce/2008/msg00001.html>
>> 
>> Hm. Must have missed that one. Oh well, c'est la vie.
>> 
>> The servers still not broke, so I'm still not going to fix it, if that's
>> what your point is.
>> 
>> Gordon
>Indeed. Dont. However in a fit of madness  to get my dev machine running 
>what the real machine was I upgraded to etch..I have to say it just 
>...worked!

I've remotely upgraded several server to Etch and in-general, they did
just work! I could probably do the same for this, probably even without a
reboot as the kernels OK, (and it doesn't run X or anything like that),
but the chances of it still being turned on in a few months are slim -
my client's "re-structuring", so who knows what'll happen with it...

Gordon

On Mon, 29 Sep 2008 08:50:18 +0000 (UTC), Gordon Henderson
<gordon+usenet@drogon.net> wrote:

>In article <gbots9$1ckm$1@energise.enta.net>,
>Paul Cupis   wrote:
>>Gordon Henderson wrote:
>>> Debian Sarge is still being maintained, so I'm happy.
>>
>>Not by the Debian Project:
>>
>><http://lists.debian.org/debian-announce/2008/msg00001.html>
>
>Hm. Must have missed that one. Oh well, c'est la vie.
>
>The servers still not broke, so I'm still not going to fix it, if that's
>what your point is.

One of our servers is still 3.1.
ISTR doing some research and upgrading to Etch seemed fraught with
danger and/or very time-consuming.

So I'm gradually moving stuff off that box and will probably wipe and
start again. We've got CentOS on the other boxes, and I'm getting used
to it now.

I read this:
http://www.debianadmin.com/upgrade-sarge-to-etch.html
and the comments, which led me to the official advice:
http://www.debian.org/releases/etch/i386/release-notes/ch-upgrading.en.html

NASTY!

-- 
Free personal divertable Phone number: <http://www.bizorg.co.uk/personalnos.htm>

In article ,
David Quinton   wrote:
>On Mon, 29 Sep 2008 08:50:18 +0000 (UTC), Gordon Henderson
><gordon+usenet@drogon.net> wrote:
>
>>In article <gbots9$1ckm$1@energise.enta.net>,
>>Paul Cupis   wrote:
>>>Gordon Henderson wrote:
>>>> Debian Sarge is still being maintained, so I'm happy.
>>>
>>>Not by the Debian Project:
>>>
>>><http://lists.debian.org/debian-announce/2008/msg00001.html>
>>
>>Hm. Must have missed that one. Oh well, c'est la vie.
>>
>>The servers still not broke, so I'm still not going to fix it, if that's
>>what your point is.
>
>One of our servers is still 3.1.
>ISTR doing some research and upgrading to Etch seemed fraught with
>danger and/or very time-consuming.
>
>So I'm gradually moving stuff off that box and will probably wipe and
>start again. We've got CentOS on the other boxes, and I'm getting used
>to it now.
>
>I read this:
>http://www.debianadmin.com/upgrade-sarge-to-etch.html
>and the comments, which led me to the official advice:
>http://www.debian.org/releases/etch/i386/release-notes/ch-upgrading.en.html

>NASTY!

Yup. the few I have upgraded weren't live by any stretch of the
imagination ... My builds over the years have more or less been "to meet
a purpose", so build the server for what it's going to do, then leave it
to it (apart from security issues)... I build most for an active life of
3 years with the aim to replce by the end of their 4th year.

Gordon

Gordon Henderson wrote:
> In article ,
> David Quinton   wrote:
>> On Mon, 29 Sep 2008 08:50:18 +0000 (UTC), Gordon Henderson
>> <gordon+usenet@drogon.net> wrote:
>>
>>> In article <gbots9$1ckm$1@energise.enta.net>,
>>> Paul Cupis   wrote:
>>>> Gordon Henderson wrote:
>>>>> Debian Sarge is still being maintained, so I'm happy.
>>>> Not by the Debian Project:
>>>>
>>>> <http://lists.debian.org/debian-announce/2008/msg00001.html>
>>> Hm. Must have missed that one. Oh well, c'est la vie.
>>>
>>> The servers still not broke, so I'm still not going to fix it, if that's
>>> what your point is.
>> One of our servers is still 3.1.
>> ISTR doing some research and upgrading to Etch seemed fraught with
>> danger and/or very time-consuming.
>>
>> So I'm gradually moving stuff off that box and will probably wipe and
>> start again. We've got CentOS on the other boxes, and I'm getting used
>> to it now.
>>
>> I read this:
>> http://www.debianadmin.com/upgrade-sarge-to-etch.html
>> and the comments, which led me to the official advice:
>> http://www.debian.org/releases/etch/i386/release-notes/ch-upgrading.en.html
> 
>> NASTY!
> 
> Yup. the few I have upgraded weren't live by any stretch of the
> imagination ... My builds over the years have more or less been "to meet
> a purpose", so build the server for what it's going to do, then leave it
> to it (apart from security issues)... I build most for an active life of
> 3 years with the aim to replce by the end of their 4th year.
> 
> Gordon

I think that is what any reasonable system administrator would do: The 
only reasons I EVER upgrade a server are to address a specific issue. 
Maybe some software I want to run wont run on an earlier library or 
kernel, or maybe there is a specific security issue.

Random upgrading 'just because its there' is something for desktop 
applications only, which are always full of bugs anyway.

I still haven't a web browser that performs adequately against ALL sites..

I would say that 3 years is about right: after that it may well be 
degraded to a lesser function, or more usually, a strip down, and new 
disks and new OS goes in.

I don't regard a 24x7 machine's disks as having an effective lifetime 
beyond 3 years anyway. Bitter experience shows that if heavily used, 
that's about it. A friend who does matrix analysis on obscure 
mathematical functions has three debians running 24x7 and all swapping 
giant matrices out to temp files, and they last about 13-18 months on 
average :-(

My main server here is coming up to 3 years, and its had one upgrade, 
and now sports a second disk that is mirroring the first..because I KNOW 
that one day the primary is going to screw up. It nearly did after fluff 
and dust clogged the fans, and it crashed buggering up a few files. 
Fortunately I had an old backup somewhere else, and the files were more 
or less unchanged since then.

In article ,
The Natural Philosopher  <a@b.c> wrote:

>I would say that 3 years is about right: after that it may well be 
>degraded to a lesser function, or more usually, a strip down, and new 
>disks and new OS goes in.

Or just scrapped - however I hate doing that (don't we all!) One thing
in the favour of 'recycling' is that processor speeds haven't really
made much inroads in the past 3 years as they have in previous years -
the server that triggered this thread has dual 2.8GHz Xeons with 1MB
Caches..

>I don't regard a 24x7 machine's disks as having an effective lifetime 
>beyond 3 years anyway. Bitter experience shows that if heavily used, 
>that's about it. A friend who does matrix analysis on obscure 
>mathematical functions has three debians running 24x7 and all swapping 
>giant matrices out to temp files, and they last about 13-18 months on 
>average :-(

Buy better disks :)

Saying that, I have one remote server with a dodgy drive - an 80GB Maxtor
IDE drive - that's now 3 years old, however it's part of a RAID-1 set
and smart has only reported a few non-critical errors so-far... I've
had a couple of other Maxtors fail, so never buy them now. I have busy
busy web servers where the disks are rattling along at about 1K sector
writes/sec - mostly apache and mysql log-files though.

>My main server here is coming up to 3 years, and its had one upgrade, 
>and now sports a second disk that is mirroring the first..because I KNOW 
>that one day the primary is going to screw up. It nearly did after fluff 
>and dust clogged the fans, and it crashed buggering up a few files. 
>Fortunately I had an old backup somewhere else, and the files were more 
>or less unchanged since then.

Get lm-sensors going. Best thing I did, and hddtemp. Graph it too and
you'll see trends. Use snmp to monitor it and nagios to report on it and
ping you TXT messages when it all goes pear-shaped!

Gordon

Gordon Henderson wrote:
> In article ,
> The Natural Philosopher  <a@b.c> wrote:
> 
>> My main server here is coming up to 3 years, and its had one upgrade, 
>> and now sports a second disk that is mirroring the first..because I KNOW 
>> that one day the primary is going to screw up. It nearly did after fluff 
>> and dust clogged the fans, and it crashed buggering up a few files. 
>> Fortunately I had an old backup somewhere else, and the files were more 
>> or less unchanged since then.
> 
> Get lm-sensors going. Best thing I did, and hddtemp. Graph it too and
> you'll see trends. Use snmp to monitor it and nagios to report on it and
> ping you TXT messages when it all goes pear-shaped!
> 
> Gordon

Think this processor is too old to have that lot.

Its a recycled pentium from my mate wot upgrades every 3 years.

Its not fast.

In article ,
The Natural Philosopher  <a@b.c> wrote:
>Gordon Henderson wrote:
>> In article ,
>> The Natural Philosopher  <a@b.c> wrote:
>> 
>>> My main server here is coming up to 3 years, and its had one upgrade, 
>>> and now sports a second disk that is mirroring the first..because I KNOW 
>>> that one day the primary is going to screw up. It nearly did after fluff 
>>> and dust clogged the fans, and it crashed buggering up a few files. 
>>> Fortunately I had an old backup somewhere else, and the files were more 
>>> or less unchanged since then.
>> 
>> Get lm-sensors going. Best thing I did, and hddtemp. Graph it too and
>> you'll see trends. Use snmp to monitor it and nagios to report on it and
>> ping you TXT messages when it all goes pear-shaped!
>> 
>> Gordon
>
>Think this processor is too old to have that lot.
>
>Its a recycled pentium from my mate wot upgrades every 3 years.
>
>Its not fast.

It's the motherboard chipset that has the sensor hardware - if you've
got the time, it's worth seeing what it has - same for hddtemp - if the
drives support SMART they might have a temperature sensor.

Gordon