Bug 676904
Summary: | osmin.img is sometimes not a valid squashfs on livecd images | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Bruno Wolff III <bruno> |
Component: | livecd-tools | Assignee: | Bruno Wolff III <bruno> |
Status: | CLOSED RAWHIDE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | high | Docs Contact: | |
Priority: | medium | ||
Version: | rawhide | CC: | adam.stokes, agk, awilliam, bcl, bmarzins, bruno, davidz, dcantrell, dhuff, dlehman, dwysocha, extras-orphan, heinzm, Jasper.Hartline, jlaska, johannbg, jonathan, jreiser, katzj, kchamart, Kenny.Strawn, kevin, mbroz, mcepl, mcepl, meyering, prockai, rbergero, rdieter, rhe, satellitgo |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | All | ||
OS: | All | ||
Whiteboard: | AcceptedBlocker | ||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | 672265 | Environment: | |
Last Closed: | 2011-02-18 18:24:15 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: | |||
Bug Depends On: | 672265 | ||
Bug Blocks: | 657616 |
Description
Bruno Wolff III
2011-02-11 18:37:30 UTC
Brian remembers the xz change happening after the problem with live installs. However there may be some bug in squashfs-tools-4.2 that is triggered in special circumstances. I haven't seen squashfs.img corrupted nor has the test case (https://fedoraproject.org/wiki/QA:Testcase_squashfs-tools_compression) shown any problems when I run it. I'm going to see if I can get Kevin Fenzi to run the test case on the same machines that do the builds in case there is some arch related issue. I don't see any problems in the test script run on the spin01 compose machine: Using existing data directory. Building squashfs image using gzip compression. Parallel mksquashfs: Using 4 processors Creating 4.0 filesystem on ./test-squashfs/sq.img, block size 131072. [==============================================================================/] 232/232 100% Exportable Squashfs 4.0 filesystem, gzip compressed, data block size 131072 compressed data, compressed metadata, compressed fragments, compressed xattrs duplicates are removed Filesystem size 12354.88 Kbytes (12.07 Mbytes) 44.47% of uncompressed filesystem size (27783.54 Kbytes) Inode table size 314 bytes (0.31 Kbytes) 12.53% of uncompressed inode table size (2506 bytes) Directory table size 246 bytes (0.24 Kbytes) 37.27% of uncompressed directory table size (660 bytes) Xattr table size 54 bytes (0.05 Kbytes) 100.00% of uncompressed xattr table size (54 bytes) Number of duplicate files found 2 Number of inodes 29 Number of files 28 Number of fragments 1 Number of symbolic links 0 Number of device nodes 0 Number of fifo nodes 0 Number of socket nodes 0 Number of directories 1 Number of ids (unique uids + gids) 1 Number of uids 1 root (0) Number of gids 1 root (0) Testing unmounted extract using gzip compression. Parallel unsquashfs: Using 4 processors 28 inodes (232 blocks) to write [==============================================================================|] 232/232 100% created 28 files created 1 directories created 0 symlinks created 0 devices created 0 fifos Testing mounted image using gzip compression. Building squashfs image using lzo compression. Parallel mksquashfs: Using 4 processors Creating 4.0 filesystem on ./test-squashfs/sq.img, block size 131072. [==============================================================================-] 232/232 100% Exportable Squashfs 4.0 filesystem, lzo compressed, data block size 131072 compressed data, compressed metadata, compressed fragments, compressed xattrs duplicates are removed Filesystem size 12354.99 Kbytes (12.07 Mbytes) 44.47% of uncompressed filesystem size (27783.54 Kbytes) Inode table size 387 bytes (0.38 Kbytes) 15.44% of uncompressed inode table size (2506 bytes) Directory table size 287 bytes (0.28 Kbytes) 43.48% of uncompressed directory table size (660 bytes) Xattr table size 54 bytes (0.05 Kbytes) 100.00% of uncompressed xattr table size (54 bytes) Number of duplicate files found 2 Number of inodes 29 Number of files 28 Number of fragments 1 Number of symbolic links 0 Number of device nodes 0 Number of fifo nodes 0 Number of socket nodes 0 Number of directories 1 Number of ids (unique uids + gids) 1 Number of uids 1 root (0) Number of gids 1 root (0) Testing unmounted extract using lzo compression. Parallel unsquashfs: Using 4 processors 28 inodes (232 blocks) to write [==============================================================================|] 232/232 100% created 28 files created 1 directories created 0 symlinks created 0 devices created 0 fifos Testing mounted image using lzo compression. Building squashfs image using lzma compression. Parallel mksquashfs: Using 4 processors Creating 4.0 filesystem on ./test-squashfs/sq.img, block size 131072. [==============================================================================/] 232/232 100% Exportable Squashfs 4.0 filesystem, lzma compressed, data block size 131072 compressed data, compressed metadata, compressed fragments, compressed xattrs duplicates are removed Filesystem size 12354.84 Kbytes (12.07 Mbytes) 44.47% of uncompressed filesystem size (27783.54 Kbytes) Inode table size 277 bytes (0.27 Kbytes) 11.05% of uncompressed inode table size (2506 bytes) Directory table size 237 bytes (0.23 Kbytes) 35.91% of uncompressed directory table size (660 bytes) Xattr table size 54 bytes (0.05 Kbytes) 100.00% of uncompressed xattr table size (54 bytes) Number of duplicate files found 2 Number of inodes 29 Number of files 28 Number of fragments 1 Number of symbolic links 0 Number of device nodes 0 Number of fifo nodes 0 Number of socket nodes 0 Number of directories 1 Number of ids (unique uids + gids) 1 Number of uids 1 root (0) Number of gids 1 root (0) Testing unmounted extract using lzma compression. Parallel unsquashfs: Using 4 processors 28 inodes (232 blocks) to write [==============================================================================|] 232/232 100% created 28 files created 1 directories created 0 symlinks created 0 devices created 0 fifos Building squashfs image using xz compression. Parallel mksquashfs: Using 4 processors Creating 4.0 filesystem on ./test-squashfs/sq.img, block size 131072. [==============================================================================/] 232/232 100% Exportable Squashfs 4.0 filesystem, xz compressed, data block size 131072 compressed data, compressed metadata, compressed fragments, compressed xattrs duplicates are removed Filesystem size 12354.97 Kbytes (12.07 Mbytes) 44.47% of uncompressed filesystem size (27783.54 Kbytes) Inode table size 322 bytes (0.31 Kbytes) 12.85% of uncompressed inode table size (2506 bytes) Directory table size 282 bytes (0.28 Kbytes) 42.73% of uncompressed directory table size (660 bytes) Xattr table size 54 bytes (0.05 Kbytes) 100.00% of uncompressed xattr table size (54 bytes) Number of duplicate files found 2 Number of inodes 29 Number of files 28 Number of fragments 1 Number of symbolic links 0 Number of device nodes 0 Number of fifo nodes 0 Number of socket nodes 0 Number of directories 1 Number of ids (unique uids + gids) 1 Number of uids 1 root (0) Number of gids 1 root (0) Testing unmounted extract using xz compression. Parallel unsquashfs: Using 4 processors 28 inodes (232 blocks) to write [==============================================================================|] 232/232 100% created 28 files created 1 directories created 0 symlinks created 0 devices created 0 fifos Testing mounted image using xz compression. I checked desktop-i386-20110212.21.iso from the nightly composes and osmin.img is a valid squashfs image. I'll keep spot checking these as they come out. It's possible that whatever was causing the bad squashfs images isn't present any more. The 64-bit version http://alt.fedoraproject.org/pub/alt/nightly-composes/desktop/desktop-x86_64-20110212.21.iso boots but goes into a loop erasing and re-displaying the automatic login dialog box. Going to VT2, logging in as "liveuser" and requesting "ps" shows dozens of "pam: gdm-password" processes, each in Sleep state waiting on poll_s, and each with the same parent. Trying "export DISPLAY=:0.0; liveinst" fails, complaining 1) libpk-gtk-module.so shared library not found; 2) AT-SPI: Accessability bus not found; 3) Unknown property: GtkDialog.has-separator. So, earlier failures stand in the way of getting good information about liveinst. That's a different bug, that is being looked into separately. You can actually work around it, but it won't help as the main issue is a kernel bug that is being discussed on lkml. Ypu need to go to run level 3 and then startx. This subbug is for the issue where osmin.img was getting corrupted in some builds. We don't know what caused that so I want to keep on the lookout for more occurrences. @Bruno ... do you mind giving a quick update on this bug? Is this still an issue, what is the exposure? I'll have a better idea in a couple of hours. I have both a nightly desktop image and an image I composed last night I want to look at today. I have a meeting in a few minutes, but will probably be able to take a time out to look at them after that. If both show osmin.img as a valid squashfs I think the risk is low. Both images had osmin.img as a valid squashfs image. So while I worry a little that we don't know the cause of the bad images, I don't think this is likely to block the release even if the glitch shows up again. Per 2011-02-18 Meeting: #agreed 676904 - accepted as Alpha Blocker. Appears to be resolved with latest F15 stable kernel Since the problem hasn't show up again, I suspect this was probably a kernel issue which previously was fixed or perhaps a variant manifestation of the cow problem. |