|
From: Friedrich L. <fl...@fl...> - 2006-04-08 22:48:35
Attachments:
patch-init.d-syslog
|
Hi! I'm proposing a change to etc/init.d/ startscripts, namely adding a filesystem init routine to the start section which unpacks parts of the /var/ filesystem if the underlying filesystem which some daemon needs is currently empty. This situation will happen when eg. LV 'log' gets mounted at /var/log. Attached is an example patch to etc/init.d/sysconfig. As you can see current existing files are not overwritten. In case some files already exist but the check determines to do the umpacking some error messages are genereated by tar because of the already existing files. Only filtering stderr (2> /dev/null) does not seem logical to me as I wanted to show if something gets done, so I accepted the possible case of distracting error messages. Those only say that a specific file already existed and wasn't overwritten. I would add similar patches to those init script which I stumble upon and am able to test. -- MfG / Regards Friedrich Lobenstock ____________________________________________________________________ Friedrich Lobenstock Linux Services Lobenstock URL: http://www.lsl.at/ Email: fl...@fl... ____________________________________________________________________ |
|
From: Heiko Z. <he...@zu...> - 2006-04-10 13:47:19
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hey, On Sat, April 8, 2006 17:48, Friedrich Lobenstock wrote: > Hi! > > > I'm proposing a change to etc/init.d/ startscripts, namely adding a > filesystem init routine to the start section which unpacks parts of the > /var/ filesystem if > the underlying filesystem which some daemon needs is currently empty. This > situation will happen when eg. LV 'log' gets mounted at /var/log. > > Attached is an example patch to etc/init.d/sysconfig. > > > As you can see current existing files are not overwritten. In case some > files already exist but the check determines to do the umpacking some > error messages are genereated by tar because of the already existing > files. > > Only filtering stderr (2> /dev/null) does not seem logical to me as I > wanted to show if something gets done, so I accepted the possible case of > distracting error messages. Those only say that a specific file already > existed and wasn't overwritten. > > I would add similar patches to those init script which I stumble upon and > am able to test. It's probably the time of the day, but I ask anyway... ;-) I can't see what this buys us, compared to the way we do thinks right now. - -- Regards Heiko Zuerker http://www.devil-linux.org -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) iEYEARECAAYFAkQ6YckACgkQUcytMSbs+YXZewCeModZE029GWUtRF8hoEiFcbsq R/kAn2ttGr1cvw7g6lQ9BQ+qdZA6bQ6f =0YBr -----END PGP SIGNATURE----- |
|
From: Friedrich L. <fl...@fl...> - 2006-04-10 19:08:32
|
Heiko Zuerker wrote on 10.04.2006 15:46 MET: > Hey, > > > On Sat, April 8, 2006 17:48, Friedrich Lobenstock wrote: > >>>Hi! >>> >>> >>>I'm proposing a change to etc/init.d/ startscripts, namely adding a >>>filesystem init routine to the start section which unpacks parts of the >>>/var/ filesystem if >>>the underlying filesystem which some daemon needs is currently empty. This >>> situation will happen when eg. LV 'log' gets mounted at /var/log. >>> >>>Attached is an example patch to etc/init.d/sysconfig. >>> >>> >>>As you can see current existing files are not overwritten. In case some >>>files already exist but the check determines to do the umpacking some >>>error messages are genereated by tar because of the already existing >>>files. >>> >>>Only filtering stderr (2> /dev/null) does not seem logical to me as I >>>wanted to show if something gets done, so I accepted the possible case of >>>distracting error messages. Those only say that a specific file already >>>existed and wasn't overwritten. >>> >>>I would add similar patches to those init script which I stumble upon and >>>am able to test. > > > It's probably the time of the day, but I ask anyway... ;-) > > I can't see what this buys us, compared to the way we do thinks right now. After you've booted a fresh DL create a LV named "var" in the VG "devil-linux", format it and then reboot the system. What do you see with "ls -al /var" after the reboot? No files if I'm correct, am I not? -- MfG / Regards Friedrich Lobenstock ____________________________________________________________________ Friedrich Lobenstock Linux Services Lobenstock URL: http://www.lsl.at/ Email: fl...@fl... ____________________________________________________________________ |
|
From: Heiko Z. <he...@zu...> - 2006-04-10 19:14:54
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Mon, April 10, 2006 14:08, Friedrich Lobenstock wrote: > Heiko Zuerker wrote on 10.04.2006 15:46 MET: > >> Hey, >> >> >> >> On Sat, April 8, 2006 17:48, Friedrich Lobenstock wrote: >> >> >>>> Hi! >>>> >>>> >>>> >>>> I'm proposing a change to etc/init.d/ startscripts, namely adding a >>>> filesystem init routine to the start section which unpacks parts >>>> of the /var/ filesystem if >>>> the underlying filesystem which some daemon needs is currently >>>> empty. This situation will happen when eg. LV 'log' gets mounted at >>>> /var/log. >>>> >>>> >>>> Attached is an example patch to etc/init.d/sysconfig. >>>> >>>> >>>> >>>> As you can see current existing files are not overwritten. In case >>>> some files already exist but the check determines to do the >>>> umpacking some error messages are genereated by tar because of the >>>> already existing files. >>>> >>>> Only filtering stderr (2> /dev/null) does not seem logical to me as >>>> I >>>> wanted to show if something gets done, so I accepted the possible >>>> case of distracting error messages. Those only say that a specific >>>> file already existed and wasn't overwritten. >>>> >>>> I would add similar patches to those init script which I stumble >>>> upon and am able to test. >> >> >> It's probably the time of the day, but I ask anyway... ;-) >> >> >> I can't see what this buys us, compared to the way we do thinks right >> now. > > After you've booted a fresh DL create a LV named "var" in the VG > "devil-linux", > format it and then reboot the system. What do you see with "ls -al /var" > after the reboot? No files if I'm correct, am I not? Now that I'm finally awake, I can understand it. ;-) Then I agree with the change. - -- Regards Heiko Zuerker http://www.devil-linux.org -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) iEYEARECAAYFAkQ6rp8ACgkQUcytMSbs+YXgqACeMHpgdw4iapj+zsUKHn4HisHo tv4AnjiRSMLAOTOKi0C8oFMX5suIkvZt =zc+L -----END PGP SIGNATURE----- |