Bug 578428
| Summary: | RuntimeError: multipath can not be run | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 6 | Reporter: | Hendrik Brueckner <brueckner> |
| Component: | anaconda | Assignee: | Ales Kozumplik <akozumpl> |
| Status: | CLOSED CURRENTRELEASE | QA Contact: | Release Test Team <release-test-team-automation> |
| Severity: | medium | Docs Contact: | |
| Priority: | low | ||
| Version: | 6.0 | CC: | brueckner, bugproxy, james.brown, jbroman, jstodola, jzeleny, maier, mgrf, moli, Stephen.Sanderlin |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | anaconda_trace_hash:fe388d42dd96120a1c2006e141aa6f8bc4cfbf0a0c586c1cc95e99f6446d3aee | ||
| Fixed In Version: | anaconda-13.21.25-1 | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2010-11-10 19:41:59 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Attachments: | |||
Created attachment 403689 [details]
Attached traceback automatically from anaconda.
backtrace only happened after sigchild from dasdfmt, so this is actually two bugs in one: 1) multipath devices were not detected 2) sigchild of dasdfmt not handled This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux major release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Major release. This request is not yet committed for inclusion. Can you also attach /tmp/syslog to this bug report? I don't know why it was left out of the traceback. Created attachment 403984 [details]
Attached traceback automatically from anaconda.
Created attachment 403988 [details]
Logfiles from the /tmp directory
Hm, there's no /tmp/syslog either. This is very strange. Where are you seeing the dasdfmt segfault? Perhaps there's some more useful information in the same place. Is this machine available for someone to look around on? Could you check whether anaconda creates /tmp/syslog at all? I used the RHEL6.0-20100328.1 build which includes version 13.21.20.2 of ananconda. The second traceback was created when configuring zfcp multipath. Yes, there should definitely be a /tmp/syslog. Do you not see one on your installation? I really need to be able to see its contents in order to rule out a variety of strange error conditions. Chris, I did not found any syslog file in /tmp. Further, the syslog daemon was not running. Ales - could you investigate why there's no syslogd running and once that's solved, maybe we can get to the bottom of this bug report. Thanks. Isn't it like this (note that on RHEL6 there's no rsyslog): The RHEL6 init.c has doklog() method which forks itself into a very simple syslog daemon reading from the /dev/klog and writing out to /tmp/syslog. But we only use init.c on everything but s390. s390 has linuxrc.s390 instead and there's no similar functionality in linuxrc.s390 as far as I can see---I could be wrong. If that's really the case, we can either merge my rsyslog changes from rawhide (we already decided this was risky in January) or I could extract out the doklog() method into a separate executable which we could call from the s390 script. Asking dcantrell. David, am I right about the linuxrc.s390 not starting any syslogd-like functionality? And should I go for the fix I propose? Ales Ales, You are correct. We do not have a logging daemon on s390x that starts from linuxrc.s390. We've never had that, actually. It's just a current limitation of having the linuxrc.s390 script. We do get logging once anaconda starts, and linuxrc.s390 is written to be absurdly verbose when it comes to failures. Since multipath runs from anaconda, we do get the traceback info, there will just be no syslog. From the looks of the traceback, it appears as though the 'multipath' command is just missing from install.img since it fails the X_OK test on line 79 in iutil.py. Sounds like we need to figure out why that's not ending up in the install root (or why it's not executable). I do not want to break out doklog() in to a separate executable that can be spawned from linuxrc.s390. We've never had that in the past and I do not want to change that now. Hendrik, can you please retest with: updates=http://akozumpl.fedorapeople.org/updates_bz578428.img and see if that helps you get any further? Thanks. Ales Fixed by 182740948aee75cc5b6e85cd01c20ce9bcd0148c, thanks to pjones, released with anaconda-13.21.25-1. (In reply to comment #14) > You are correct. We do not have a logging daemon on s390x that starts from > linuxrc.s390. We've never had that, actually. It's just a current limitation > of having the linuxrc.s390 script. > I do not want to break out doklog() in to a separate executable that can be > spawned from linuxrc.s390. We've never had that in the past and I do not want > to change that now. I once spun a custom initrd.img last year including a working rsyslogd started by linuxrc.s390 for debugging NetworkManager on s390 when it gets started by loader and before stage2 is active. Would that be an option? After all there are daemons whose syslog messages cannot be seen until anaconda runs. (In reply to comment #16) > Fixed by 182740948aee75cc5b6e85cd01c20ce9bcd0148c, thanks to pjones, released > with anaconda-13.21.25-1. This fixes iutil.exec of the multipath tool. What about the sigchild issues? Steffen, I couldn't see anything about an unhandled sigchld in the logs, that's why I asked Hendrik to retest. Hendrik, how did you find out about the unhandled SIGCHLD? Thanks. Ales ------- Comment From kumarr.com 2010-04-22 18:34 EDT------- Mirror to IBM Created attachment 409804 [details]
Attached traceback automatically from anaconda.
Created attachment 411365 [details]
Attached traceback automatically from anaconda.
Johan, the log you attached shows anaconda is in version 13.21.20.5, whereas the fix is only available in anaconda-13.21.25-1, thus the error still. About your question: yes, feel free to change architecture to 'all', this bug certainly is not architecture dependent. Thank you. Ales Created attachment 412371 [details]
Attached traceback automatically from anaconda.
------- Comment From lnx1138.ibm.com 2010-06-03 10:36 EDT------- --- ------- Comment From lnx1138.ibm.com 2010-06-03 13:51 EDT------- *** Bug 64483 has been marked as a duplicate of this bug. *** ------- Comment From edpollar.com 2010-06-16 12:02 EDT------- reassigning QA... ------- Comment From htengshe.com 2010-06-17 02:33 EDT------- Hi I am unable to install RHEL6.0 prebeta on SCSI multipath. In "Devices" screen, I add a ZFCP lun.n I add the same lun I get unhandled execption when I try to add the same lun with same WWPN but using different adapter. (ie while trying to create multipath). Attached is Exception report. Created attachment 424690 [details]
Exception Report while creating multipath RHEL6.0 prebeta
------- Comment (attachment only) From htengshe.com 2010-06-17 02:36 EDT-------
Hi, what you see in comment 31 and comment 32 is the same issue as described in bug 597223 - TypeError: list indices must be integers, not str Try to specify your zFCP multipath device in CMS config file instead of adding devices in GUI This issue is no longer reproducible, tested on x86_64 and s390x using multipath zFCP LUN. anaconda-13.21.56-1.el6, RHEL6.0-20100701.3 Moving to VERIFIED. ------- Comment From thorsten.diehl.com 2010-07-16 09:57 EDT------- Issue is still reproducable on snapshot 7 with anaconda-13.21.56-1.el6 Waiting for fix described in https://bugzilla.redhat.com/show_bug.cgi?id=597223 ------- Comment From htengshe.com 2010-08-16 04:53 EDT------- Tested on RHEl6.0 Snap10. I am able to configure multipath. However if I select only multipath device for installation, I am unable to proceed as it throws error "No usable disks have been found". The corrsponding bug for this problem is #66395. > The corrsponding bug for this problem is #66395. LTC bug 66395 - RIT 1232563 - [RHEL6 Snap9]Can't install RHEL6 Snap9 on romeolp1 with multipath disks via VNC for 'No usable disks have been found' is RH bug 622551 which is currently closed dup to RH bug 596468. Red Hat Enterprise Linux 6.0 is now available and should resolve the problem described in this bug report. This report is therefore being closed with a resolution of CURRENTRELEASE. You may reopen this bug report if the solution does not work for you. ------- Comment From kumarr.com 2010-04-22 18:34 EDT------- |
The following was filed automatically by anaconda: anaconda 13.21.20.2 exception report Traceback (most recent call first): File "/usr/lib64/python2.6/subprocess.py", line 1092, in _execute_child raise child_exception File "/usr/lib64/python2.6/subprocess.py", line 595, in __init__ errread, errwrite) File "/usr/lib/anaconda/iutil.py", line 134, in execWithRedirect env=env) File "/usr/lib/anaconda/storage/devices.py", line 3109, in setup stderr = "/dev/tty5") File "/usr/lib/anaconda/storage/devicetree.py", line 1929, in populate mp.setup() File "/usr/lib/anaconda/storage/__init__.py", line 374, in reset self.devicetree.populate() File "/usr/lib/anaconda/storage/__init__.py", line 109, in storageInitialize storage.reset() File "/usr/lib/anaconda/dispatch.py", line 205, in moveStep rc = stepFunc(self.anaconda) File "/usr/lib/anaconda/dispatch.py", line 126, in gotoNext self.moveStep() File "/usr/lib/anaconda/gui.py", line 1308, in nextClicked self.anaconda.dispatch.gotoNext() RuntimeError: multipath can not be run