When installing Florence Beta3 (Fisher) on Perc3/QC and Perc3/DC, the installer fails with "...IOError: [Errno 28] No space left on device..." message. This problem is independent of the container size (tried 9GB to 34GB containers) on PE1400(Gecko II), PE120L (Saber), and PE4400 (Opal). Perc2/DC works ok.
Assigning to a developer.
there should be a full traceback with this error. Could you attach it to the bug report, please?
Following are the configurations: 1) - PE2400 + 5 9Gb HDDs + Perc3/DC ---> RAID 5 2) - PE1400 + 2 36Gb HDDs + Perc3/DC ---> RAID 1 3) - PE4400 + 5 36Gb HDDs + Perc3/DC ---> RAID 1 & 5 We were able to install on Perc3/QC with no problem.
This traceback was provided by email from Tesfamariam_Michael, thanks! Traceback (innermost last): File "/var/tmp/anaconda-7.1//usr/lib/anaconda/iw/progress_gui.py", line 19, in run rc = self.todo.doInstall () File "/var/tmp/anaconda-7.1//usr/lib/anaconda/todo.py", line 1645, in doInstall (p, self.intf.messageWindow, pkgTimer)) File "/var/tmp/anaconda-7.1//usr/lib/anaconda/todo.py", line 1390, in instCallback self.instLog.flush () IOError: [Errno 28] No space left on device Local variables in innermost frame: pkgTimer: <timer.Timer instance at 95c7c68> param: (<progress_gui.InstallProgressWindow instance at 84757e0>, <method InstallInterface.messageWindow of InstallInterface instance at 820d098>, <timer.Timer instance at 95c7c68>) what: 2 total: 0 self: <todo.ToDo instance at 84593c8> h: <header object at 88cfe68> amount: 0 intf: <progress_gui.InstallProgressWindow instance at 84757e0> .10: (<progress_gui.InstallProgressWindow instance at 84757e0>, <method InstallInterface.messageWindow of InstallInterface instance at 820d098>, <timer.Timer instance at 95c7c68>) messageWindow: <method InstallInterface.messageWindow of InstallInterface instance at 820d098>
just to clarify, how large were your / and /tmp filesystems? Were either unusually small? (The traceback is trying to write the install log to /tmp/install.log when it ran out of space...)
Initially this is what we had: /boot 32MB /usr 4GB /var 4GB swap 512MB /home 4GB / --grow (this is more than enough) Another one we tried is: /boot 32MB swap 512MB / --grow
What is the firwmare version of the RAID controller? Was the installation failing with PE6300 systems?
switch to vc2 after the traceback and type "df" and send the output.
I have tried all firmmare 1.57 down to 1.43.
We need a copy of this hardware here (in-house) for analysis; please ship us one so we can analyze this problem.
Do you still see this problem in RC2?
No, I don't see this issue anymore.
closing
This issue was resolved in RC2 but it is re-occuring in qa0319 and qa0322 releases. This time it happens on PowerEdge 1550(Saber) with PERC/3DCL (Taos Lite). It doesn't occure in PowerEdge 1400 anymore. When installing the latest releases on PE1550 with PERC3/DCL (two 16GB of RAID1 firmware 1.57), the installer terminates abnormally with signal 11 as soon anaconda starts. I will test if this occures with PERC3/DC (Taos) tomorrow.
there was a kudzu bug causing segfaults. I'd say this is a completely different bug than what you saw earlier with the out of space bug.
I'm able to install without a problem on the 1550 here with a PERC/3DC installed.
I have been able to install with QA0327 (kernel 2.4.2-0.1.40) on PowerEdge 1550 with PERC3/DCL (Taos Lite). This is the same configuration that failed with the last two builds (QA0319 & QA0322). The issue was occuring only on PERC3/DCL which is a similar to PERC3/DC but without the battery and charger. I will send you a tool that lets you switch from one to the other.
you agree that it is working OK now?
Yes, I agree.
excellent, closing.