practical limit on the number of UIDs

practical limit on the number of UIDs

Tzafrir Cohen tzafrir at cohens.org.il
Wed Jun 29 16:57:08 IDT 2011


On Wed, Jun 29, 2011 at 02:38:47PM +0300, Tom Goren wrote:
> Also, there is another issue:
> 
> 1 million users could be a lot to lose, and a few files aren't as resilient
> as a DB, which gives you things like replication and redundancy.
> 
> Since you stated that only a few users will connect at a time, then sure,
> performance isn't an issue.
> 
> Maintainability, like Shachar stated is definitely an issue at that scale.
> 
> I just wanted to add that data reliability is a factor when dealing with
> such a large data set (even just using the 'passwd' command can bork a
> shadow file on occasion).

"Users in a Linux system" does not imply "lines in passwd". NSS and PAM
can take users from different sources. LDAP was mentioned, and there are
certainly other options.

PAM and NSS are completely userspace (right?). Should the kernel even
care about users that are not currently logged on?

-- 
Tzafrir Cohen         | tzafrir at jabber.org | VIM is
http://tzafrir.org.il |                    | a Mutt's
tzafrir at cohens.org.il |                    |  best
tzafrir at debian.org    |                    | friend



More information about the Linux-il mailing list