Re: [2008.x] 'opkg upgrade' interrupted.
Joel Newkirk wrote: [...] The short solution - 'opkg update' and 'opkg install dropbear' from terminal, then 'opkg upgrade' via ssh. opkg is imperfect... Actually I believe it's dropbear's fault. Kind regards, -- Lech Karol Pawłaszek ike You will never see me fall from grace [KoRn] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [2008.x] 'opkg upgrade' interrupted.
On Sun, 23 Nov 2008 17:55:06 +0100, Lech Karol Pawłaszek [EMAIL PROTECTED] wrote: Joel Newkirk wrote: [...] The short solution - 'opkg update' and 'opkg install dropbear' from terminal, then 'opkg upgrade' via ssh. opkg is imperfect... Actually I believe it's dropbear's fault. Kind regards, The snafu of killing your active SSH during upgrade is not strictly Dropbear's fault, more the fault of the upgrade/install scripts accompanying the package. Regardless, I was unclear, but was referring to the other 'issues' when I referred to opkg as 'imperfect' (being polite). Unless it's in the middle of a string of inter-related dependencies, there's no reason that previously-handled packages should be processed again. And ideally, 'environment-breaking' actions like restarting dropbear or restarting X should be deferred until all package upgrades are completed. j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
[2008.x] 'opkg upgrade' interrupted.
Hey all. I'm was just installing a stable 2008.9 image. The only changes made is that a terminal is installed and the /etc/opkg/* is changed to pointing to the testing repository instead. Then, from the local terminal, I run opkg update and then opkg upgrade. The upgrade is working for some good 20-30 minutes downloading and installing new packages until X is shut down. Apparently opkg wanted to restart X or something? Then I figure I'll do it over ssh instead, so I log in and run 'opkg upgrade' again. To my surprise I see that it starts all over again, downloading and installing (?) the same packages over again. Didn't the previous upgrade just do that? Are the downloaded and installed files stored in a temporary directory in the meantime? I checked the used disk-space and found it to be only 81-82 MB. Didn't check how much was used before the first run though. Or does the opkg wait with updating its database of installed packages until the complete upgrade has finished? Which means I by now actually have the latest and greatest, just opkg doesn't know it is? Anyway, the upgrade via ssh failed miserably as well. Of course dropbear had to be restarted, I lost connection and opkg was interrupted (again). Next try is screen, which is currently running, downloading and installing the exact same packages done by the previous two upgrades. Is there some way to make opkg continue where interrupted or some way of specifying I don't want a certain service restarted (like dropbear) ? -Ivar Mossin ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
[2008.x] 'opkg upgrade' interrupted.
One possible way round this is to run your upgrade within a screen session. The upgrade will continue even if you get disconnected from the session, in which case you can reconnect later. That's what I did just now. Although I was a bit scared when I saw this: Configuring dropbear System startup links for /etc/init.d/dropbear already exist. Restarting Dropbear SSH server: Connection to 192.168.0.202 closed by remote host. Connection to 192.168.0.202 closed. [EMAIL PROTECTED]:~$ ssh [EMAIL PROTECTED] [EMAIL PROTECTED]'s password: [EMAIL PROTECTED]:~# screen -ls No Sockets found in /tmp/screens/S-root. ... I don't know why, but somehow I did actually try 'screen -R' anyway, even though I saw the list was empty. And it did actually work. :) I wouldn't call this a solution though, just a work-around. -Ivar Mossin ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: [2008.x] 'opkg upgrade' interrupted.
On Sat, 22 Nov 2008 23:46:38 +0100, Ivar Mossin [EMAIL PROTECTED] wrote: Hey all. I'm was just installing a stable 2008.9 image. The only changes made is that a terminal is installed and the /etc/opkg/* is changed to pointing to the testing repository instead. Then, from the local terminal, I run opkg update and then opkg upgrade. The upgrade is working for some good 20-30 minutes downloading and installing new packages until X is shut down. Apparently opkg wanted to restart X or something? Then I figure I'll do it over ssh instead, so I log in and run 'opkg upgrade' again. To my surprise I see that it starts all over again, downloading and installing (?) the same packages over again. Didn't the previous upgrade just do that? Are the downloaded and installed files stored in a temporary directory in the meantime? I checked the used disk-space and found it to be only 81-82 MB. Didn't check how much was used before the first run though. Or does the opkg wait with updating its database of installed packages until the complete upgrade has finished? Which means I by now actually have the latest and greatest, just opkg doesn't know it is? Anyway, the upgrade via ssh failed miserably as well. Of course dropbear had to be restarted, I lost connection and opkg was interrupted (again). Next try is screen, which is currently running, downloading and installing the exact same packages done by the previous two upgrades. Is there some way to make opkg continue where interrupted or some way of specifying I don't want a certain service restarted (like dropbear) ? -Ivar Mossin The short solution - 'opkg update' and 'opkg install dropbear' from terminal, then 'opkg upgrade' via ssh. opkg is imperfect... j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community