[Fwd: [Bug 154763] Grub is inadequate without Lilo for backup]

Mike Bird mgb-fedora at yosemite.net
Wed Apr 20 08:45:53 UTC 2005


On Wed, 2005-04-20 at 01:00, Arjan van de Ven wrote:
> Now, you want more track record for grub. But would YOU run grub even if
> lilo was put back? I thought so. So how would grub get that track record
> for you? It wouldn't. Result: in our eyes it would never get that track
> record. Proof: Grub has built that track record over the past 2 1/2
> years, where initially it was a bit shaky perhaps (although I don't
> think it was that bad) but is quite mature nowadays. You just failed to
> notice I guess because you assumed it was not reliable and thus didn't
> use it.

We've run Grub on some systems (not many) because some techs and/or
friends setup there personal workstations and/or home systems with
default workstation installs.  Some still survive and some had to be
rescued.  The most recent IIRC was an FC3 system a few months ago. 
Others on this list have reported recent reliability problems with SATA.

A lot of people have been bitten by Grub's inability to handle MD.  Grub
now claims to handle MD but that code does NOT have 2 1/2 years of
proven reliability because it was only written a month ago.  We can try
it on test boxes and then a few workstations and then minor servers
etc.  Maybe in 2 1/2 years we can consider removing Lilo.

The ONLY time I've ever had a problem with Lilo is when I made a mistake
- omitting "linear" from the conf file.  Lilo has been used on most
systems here for roughly ten years.  I like software that's more
reliable than me.

You don't have to use Lilo if you don't want to.  You don't have to make
Lilo obvious for newbies.  Why all the fuss?  How does it hurt you if
serious sysadmins use Lilo?

--Mike Bird




More information about the test mailing list