Re: unique IDs for connections to ipsvd programs
Thilo Bangert <[email protected]>
| Newsgroups | gmane.comp.misc.pape.general |
|---|---|
| Message-ID | <[email protected]> |
Lasse Kliemann <[email protected]> said: > * Message by -Charlie Brady- from Tue 2007-11-06: > > On Sat, 13 Oct 2007, Lasse Kliemann wrote: > >> I have a suggestion for another environment variable to be set by > >> the ipsvd > >> programs. How about generating a unique ID for each connection and > >> passing this ID on to 'prog' via an environment variable and also > >> writing the ID to > >> the log? > > > > Are PIDs not "unique enough" for log correlation purposes? > > If one has a rough estimate at which time the connection in question > took place, then perhaps. But not given that, I find PIDs not unique > enough. but when you pipe your log through multilog, you will have a timestamp with microsecond resolution, no? granted, i don't think i fully understand why PID's are not enough. for the lifetime of the connection they are obviosly unique... > > Actually, I use PIDs for that purpose right now (in combination with > the client port number), and I find the solution unsatisfactory > (although sufficient for the moment).
signature.asc
(application/pgp-signature, 189 B)
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (GNU/Linux) iD8DBQBHM0HjxRElEoA5AncRAi1PAJ9gNaj10N/s3d6cfo0Hr5KNp3cWggCgn/S3 ttkWzS1H4TMNySnWn21EpXY= =Hd2L -----END PGP SIGNATURE-----