Created attachment 397067 [details] /tmp/*log.tgz Description of problem: Version-Release number of selected component (if applicable): * anaconda-13.32 How reproducible: * 1 out of 1 attempts on *bare* metal (not virt) Steps to Reproduce: 1. Follow instructions at https://fedoraproject.org/wiki/QA/TestCases/InstallSourceCdrom Actual results: [anaconda.log] 08:38:20,984 INFO anaconda: switching from media #1 to #2 for Packages/system-config-lvm-1.1.12-2.fc13.noarch.rpm 08:38:25,894 DEBUG anaconda: isys.py:umount()- going to unmount /mnt/source, removeDir = False [storage.log] 08:38:25,893 DEBUG storage: found OpticalDevice instance (0x7f9d03611790) -- name = sr0 status = True parents = [] kids = 0 id = 4 uuid = None format = <storage.formats.fs.Iso9660FS object at 0x7f9d03611810> size = 691.609375 major = 11 minor = 0 exists = True sysfs path = /devices/pci0000:00/0000:00:06.0/host0/target0:0:0/0:0:0:0/block/sr0 partedDevice = <parted.device.Device object at 0x7f9d03611850> target size = 0 path = /dev/sr0 format args = [] 08:38:26,000 DEBUG storage: OpticalDevice.eject: sr0 ; status: True ; 08:38:26,005 DEBUG storage: OpticalDevice.teardown: sr0 ; status: True ; 08:38:26,112 WARNING storage: error ejecting cdrom sr0: (5, 'Input/output error') Expected results: Ejected disc Additional info: * See attached log files -rw-r--r-- root/root 31466 2010-03-01 08:39 tmp/X.log -rw-r--r-- root/root 13278 2010-03-01 08:38 tmp/anaconda.log -rw-r--r-- root/root 57716 2010-03-01 08:38 tmp/program.log -rw-r--r-- root/root 183168 2010-03-01 08:38 tmp/storage.log -rw-r--r-- root/root 80675 2010-03-01 08:39 tmp/syslog -rw-r--r-- root/root 136712 2010-03-01 08:22 tmp/yum.log
= Additional information = From tty2, I can successfully eject the disc by typing: # eject /dev/cdrom
= Tested workaround = 1) change to tty2 2) type: eject /dev/cdrom 3) insert next disc 4) type: mount -o ro /dev/cdrom /mnt/source 5) Return to tty6 6) Select OK to proceed with install NOTE: step#4 was required, for some reason anaconda wouldn't continue the install unless I manually mounted the next ISO
What about if you flip to tty2 and instead: PYTHONPATH=/usr/lib/anaconda python >>> import _isys >>> import os >>> fd = os.open("/dev/sr0", os.RD_ONLY|os.NONBLOCK) >>> _isys.ejectcdrom(fd)
Figures, when retesting ... the problem did not occur. I'll continue testing ...
Reproduced, the problem surfaces only when you mediacheck the disc in loader. The suggestion in comment#3 does not resolve the issue. >>> _isys.ejectcdrom(fd) Traceback (most recent call last): File "<stdin>", line 1, in <module> SystemError: (5, 'Input/output error')
Well, the point of the suggestion in comment #3 was to isolate that it was the same code anaconda's using that causes the problem without involving all of anaconda.
Moving to F13Beta per point#3 of the Beta release criteria (https://fedoraproject.org/wiki/Fedora_13_Beta_Release_Criteria#Beta_Release_Requirements)
Documented known issue and workaround for Alpha users at https://fedoraproject.org/wiki/Common_F13_bugs#failure-to-eject-cdrom
James, I've taken a quick look at what the eject command does differently, Can you try some python code on tty2 again, but this time this one: PYTHONPATH=/usr/lib/anaconda python >>> import _isys >>> import os >>> fd = os.open("/dev/sr0", os.O_RDWR|os.O_NONBLOCK) >>> _isys.ejectcdrom(fd) And can you also run eject in verbose mode: eject -v /dev/sr0 Thanks, Hans
Hans - I already asked him to do that. Please see comment #3 and comment #5.
(In reply to comment #10) > Hans - I already asked him to do that. Please see comment #3 and comment #5. Chris, Look closer: 1) My python code has other open flags (the same as the eject binary uses) 2) The eject command I gave has a "-v" James, Could you please provide the info I requested in comment #9. Note that it would be good to have the output of both when booted with and without mediacheck. Thanks, Hans
(In reply to comment #11) > (In reply to comment #10) > > Hans - I already asked him to do that. Please see comment #3 and comment #5. > > Chris, > > Look closer: > 1) My python code has other open flags (the same as the eject binary uses) > 2) The eject command I gave has a "-v" > > James, > > Could you please provide the info I requested in comment #9. Note that it would > be good to have the output of both when booted with and without mediacheck. Due to further hal-ectomy changes during the F-13-Alpha release process, the bits released as F-13-Alpha no longer have the 'eject' binary available. I have copied the /usr/sbin/eject binary from the F13 eject package and run it on the system. = with mediacheck = # eject -v -v -v /dev/cdrom /tmp/eject: device name is `/dev/cdrom' /tmp/eject: expanded name is `/dev/cdrom' /tmp/eject: `/dev/cdrom' is a link to `/dev/sr0' /tmp/eject: `/dev/sr0' is mounted at `/mnt/source' /tmp/eject: checking if device "/dev/sr0" has a removable or hotpluggable flag /tmp/eject: unmounting device `/dev/sr0' from `/mnt/source' /tmp/eject: `/dev/sr0' is not a multipartition device /tmp/eject: trying to eject `/dev/sr0' using CD-ROM eject command /tmp/eject: CD-ROM eject command failed /tmp/eject: trying to eject `/dev/sr0' using SCSI commands /tmp/eject: SCSI eject succeeded = without mediacheck = # eject -v -v -v /dev/cdrom /tmp/eject: device name is `/dev/cdrom' /tmp/eject: expanded name is `/dev/cdrom' /tmp/eject: `/dev/cdrom' is a link to `/dev/sr0' /tmp/eject: `/dev/sr0' is not mounted /tmp/eject: unable to open /etc/fstab: No such file or directory /tmp/eject: checking if device "/dev/sr0" has a removable or hotpluggable flag /tmp/eject: `/dev/sr0' is not a multipartition device /tmp/eject: trying to eject `/dev/sr0' using CD-ROM eject command /tmp/eject: CD-ROM eject command succeeded
Moving this over to pjones as discussed on irc.
Retested and confirmed the reported behavior remains present in F-13-Beta-TC1
anaconda-13.37-1.fc13 has been submitted as an update for Fedora 13. http://admin.fedoraproject.org/updates/anaconda-13.37-1.fc13
anaconda-13.37-1.fc13 has been pushed to the Fedora 13 stable repository. If problems still persist, please make note of it in this bug report.