Bug 1744354 - mxfs.xfs fails with 'agsize (1024 blocks) too small, need at least 4096 blocks' in openQA 'no swap' and 'software RAID' install tests
Summary: mxfs.xfs fails with 'agsize (1024 blocks) too small, need at least 4096 block...
Keywords:
Status: CLOSED DUPLICATE of bug 1743753
Alias: None
Product: Fedora
Classification: Fedora
Component: anaconda
Version: rawhide
Hardware: x86_64
OS: Linux
unspecified
high
Target Milestone: ---
Assignee: Vendula Poncova
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard: openqa
Depends On:
Blocks: F31FinalBlocker
TreeView+ depends on / blocked
 
Reported: 2019-08-21 22:23 UTC by Adam Williamson
Modified: 2019-08-22 15:00 UTC (History)
7 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2019-08-22 15:00:32 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)
anaconda.log (54.97 KB, text/plain)
2019-08-21 22:29 UTC, Adam Williamson
no flags Details
dbus.log (2.92 KB, text/plain)
2019-08-21 22:30 UTC, Adam Williamson
no flags Details
program.log (43.92 KB, text/plain)
2019-08-21 22:30 UTC, Adam Williamson
no flags Details
storage.log (258.31 KB, text/plain)
2019-08-21 22:30 UTC, Adam Williamson
no flags Details
syslog (691.25 KB, text/plain)
2019-08-21 22:31 UTC, Adam Williamson
no flags Details

Description Adam Williamson 2019-08-21 22:23:14 UTC
Since Fedora-Rawhide-20190816.n.0 - and including Fedora-31-20190820.n.4, so this is on Rawhide and F31 both - the openQA 'install_no_swap' and 'install_software_raid' tests are failing consistently on every compose. The failure seems to be when formatting the root partition, as XFS because the tests run from the Server DVD:

03:23:13,335 INF program: Running... mkfs.xfs -f /dev/mapper/fedora_ibm--p8--kvm--03--guest--02-00
03:23:13,347 INF program: stderr:
03:23:13,347 INF program: b'agsize (1024 blocks) too small, need at least 4096 blocks'
03:23:13,348 INF program: b'Usage: mkfs.xfs'
03:23:13,348 INF program: b'/* blocksize */\t\t[-b size=num]'
03:23:13,348 INF program: b'/* metadata */\t\t[-m crc=0|1,finobt=0|1,uuid=xxx,rmapbt=0|1,reflink=0|1]'
03:23:13,348 INF program: b'/* data subvol */\t[-d agcount=n,agsize=n,file,name=xxx,size=num,'
03:23:13,348 INF program: b'\t\t\t    (sunit=value,swidth=value|su=num,sw=num|noalign),'
03:23:13,348 INF program: b'\t\t\t    sectsize=num'
03:23:13,348 INF program: b'/* force overwrite */\t[-f]'
03:23:13,349 INF program: b'/* inode size */\t[-i perblock=n|size=num,maxpct=n,attr=0|1|2,'
03:23:13,349 INF program: b'\t\t\t    projid32bit=0|1,sparse=0|1]'
03:23:13,349 INF program: b'/* no discard */\t[-K]'
03:23:13,349 INF program: b'/* log subvol */\t[-l agnum=n,internal,size=num,logdev=xxx,version=n'
03:23:13,349 INF program: b'\t\t\t    sunit=value|su=num,sectsize=num,lazy-count=0|1]'
03:23:13,349 INF program: b'/* label */\t\t[-L label (maximum 12 characters)]'
03:23:13,349 INF program: b'/* naming */\t\t[-n size=num,version=2|ci,ftype=0|1]'
03:23:13,349 INF program: b'/* no-op info only */\t[-N]'
03:23:13,350 INF program: b'/* prototype file */\t[-p fname]'
03:23:13,350 INF program: b'/* quiet */\t\t[-q]'
03:23:13,350 INF program: b'/* realtime subvol */\t[-r extsize=num,size=num,rtdev=xxx]'
03:23:13,350 INF program: b'/* sectorsize */\t[-s size=num]'
03:23:13,350 INF program: b'/* version */\t\t[-V]'
03:23:13,350 INF program: b'\t\t\tdevicename'
03:23:13,351 INF program: b'<devicename> is required unless -d name=xxx is given.'
03:23:13,351 INF program: b'<num> is xxx (bytes), xxxs (sectors), xxxb (fs blocks), xxxk (xxx KiB),'
03:23:13,351 INF program: b'      xxxm (xxx MiB), xxxg (xxx GiB), xxxt (xxx TiB) or xxxp (xxx PiB).'
03:23:13,351 INF program: b'<value> is xxx (512 byte blocks).'
03:23:13,351 DBG program: Return code: 1

only those two tests are affected, for some reason (presumably something to do with the auto-generated filesystem layout - it may be to do with the root partition size, in fact, as it'll likely be bigger in those two tests than any of the others, now I come to think about it). I'll attach all other logs too.

anaconda changed in the compose in question, from anaconda-31.22-1.fc31 to anaconda-32.1-1.fc32. pykickstart went from pykickstart-3.20-4.fc31 to pykickstart-3.21-1.fc32 . pyparted, libblockdev, blivet and xfsprogs did not change.

Comment 1 Adam Williamson 2019-08-21 22:24:14 UTC
Proposing as an F31 Final blocker as a violation of "The installer must be able to create and install to any workable partition layout using any file system and/or container format combination offered in a default installer configuration."

Comment 2 Adam Williamson 2019-08-21 22:29:57 UTC
Created attachment 1606764 [details]
anaconda.log

Comment 3 Adam Williamson 2019-08-21 22:30:16 UTC
Created attachment 1606765 [details]
dbus.log

Comment 4 Adam Williamson 2019-08-21 22:30:33 UTC
Created attachment 1606766 [details]
program.log

Comment 5 Adam Williamson 2019-08-21 22:30:55 UTC
Created attachment 1606767 [details]
storage.log

Comment 6 Adam Williamson 2019-08-21 22:31:26 UTC
Created attachment 1606768 [details]
syslog

Comment 7 Vendula Poncova 2019-08-22 13:24:20 UTC
It might be a duplicate of 1743753.

Comment 8 Adam Williamson 2019-08-22 15:00:32 UTC
Yeah, they look the same indeed. You can see the same error in program.log.

*** This bug has been marked as a duplicate of bug 1743753 ***


Note You need to log in before you can comment on or make changes to this bug.