Opened 16 years ago

Closed 15 years ago

#84 closed defect (fixed)

error in sipb-xen-lvcopy

Reported by: price Owned by: price
Priority: major Milestone: Public Beta
Component: autoinstallers Version:
Keywords: Cc:

Description (last modified by quentin)

From: Alex Schwendner <alexrs@MIT.EDU>                                                                        
Subject: Error creating a SIPB Xen VM                                                                         
Date: Fri, 9 May 2008 12:22:01 -0400                                                                          


I got an error while trying to create a VM with SIPB's alpha VM
service as was asked to email sipb-xen. I tried to create a machine
with the following parameters:

Name: monoid
Memory: 256 (MiB)
Disk: 10 (GiB)
Clone Image = YES
Boot CD: None
Owner: alexrs

Here is the error message:

ERROR on remctl in operation create. This shouldn't happen! Please
email to explain how it happened. Stderr:

Error 1 on remctl ('web', 'lvcopy', 'ice3', 'monoid', 'password') :
4096+0 records in
4096+0 records out
4294967296 bytes (4.3 GB) copied, 89.8694 seconds, 47.8 MB/s
Traceback (most recent call last):
  File "/usr/sbin/sipb-xen-lvcopy", line 82, in <module>
  File "/usr/sbin/sipb-xen-lvcopy", line 73, in main
    duplicate_by_vm(*args, **kwargs)
  File "/usr/sbin/sipb-xen-lvcopy", line 56, in duplicate_by_vm
    frob_copy(target_device, target, rootpw)
  File "/usr/sbin/sipb-xen-lvcopy", line 35, in frob_copy
    fs = losetup(target, 32256)
  File "/usr/sbin/sipb-xen-lvcopy", line 12, in losetup
    return p.communicate()[0].strip()
AttributeError: 'NoneType' object has no attribute 'strip'

Best Regards,
Alex Schwendner

Change History (3)

comment:1 Changed 16 years ago by ecprice

  • Description modified (diff)

In /usr/sbin/sipb-xen-lvcopy, code like

p = Popen(ls?, stdout=PIPE) print p.stdout

prints None the first time when called with remctl. It prints the right thing (an fd) when run directly, or on the second call to Popen.

comment:2 Changed 16 years ago by quentin

  • Description modified (diff)

Is this some kind of race condition where the process either ends before we call communicate or hasn't started yet when we call communicate?

comment:3 Changed 15 years ago by price

  • Resolution set to fixed
  • Status changed from new to closed

ecprice tracked down the bug in subprocess that caused this issue, and IIRC patched it.

Also, sipb-xen-lvcopy is obsolete since completing #65.

Note: See TracTickets for help on using tickets.