Why most run Microsoft, not RedHat
Les Mikesell
lesmikesell at gmail.com
Thu Apr 19 19:59:10 UTC 2007
BRUCE STANLEY wrote:
> -
>>> Today Les Mikesell did spake thusly:
>>>
>>
>> There were problems getting linux drivers to work on certain hardware
>> back then, but it was as good in terms of 'if it worked once it will
>> keep working' and if anything broke it could be fixed by replacing the
>> appropriate file. By RH7.3 it was about as solid as anything before or
>> since... I have one of those that is within a couple of months of a
>> 4-year uptime (counter has rolled twice) - and for a few years before
>> that it was only down for the reboots to update the kernel.
>>
>> *Every* windows box here has been down many times due to virus attacks
>> and the updates required to prevent more of them in that time span.
>>
>
>
> Wow! 4 years and no reboot?
>
> We have not hat that kind of success with our
> RHEL 3 Taroon 2.4.21-4ELsmp system (2 XEONS with 8GB memory).
This is running 2.4.20-18.7 on a system with 4 gigs of RAM. It's
running our DHCP, DNS, samba, internal mail services and a few internal
web sites.
> If we do not reboot at least once every 2 weeks, we start having
> memory starvation problems.
I also have dozens of machines running RH9 2.4.20-18.9bigmem #1 SMP
or 2.4.20-20.9smp #1 SMP and CentOS 3.4 running 2.4.21-27.0.4.ELsmp
that have been up more than a year. And lots of others that have only
been down for relocation or kernel updates. I do have problems with a
bug in the Sun java j2sdk1.4.2_05 JVM that leaks memory, but I just
restart the JVM periodically instead of rebooting. If you think you
have a kernel problem you should update it. If you have an application
leaking memory, you should be able to restart it instead of rebooting.
> We think that there is a memory leak problem somwhere or
> something is grabing memory for I-O buffers but not
> releasing any for furhter use.
It's normal for unused RAM to be used for file buffers but they should
be released as needed by applications. I've heard of some kernel
versions having problems with that, but they were usually quickly
replaced by updated versions.
> Our hardware/Linux support people just tell us
> to do a reboot which fixes the problem.
Are you running the latest update? I'd expect RHEL3 to match the
current 2.4.21-47.0.1.ELsmp I have on updated Centos3 boxes. And I'd
expect the support people to tell you to update if you have a problem
and an update is available.
--
Les Mikesell
lesmikesell at gmail.com
More information about the fedora-list
mailing list