PaulHowarth/Blog/2006-10

Paul's Blog Entries for October 2006

Sunday 1st October 2006

Printer Sharing

I set up printer sharing for our Samsung ML-4500 using CUPS for the Linux clients (easy) and Samba for the Windows clients. I've always found getting Point'n'Print working for the Windows clients to be tricky, so this time I've documented the procedure at PrinterSharingWithCupsAndSamba so it'll be easy if I ever have to do it again.

Leon Gets a Haircut

leon-haircut.jpg

Mummy did it.

Monday 2nd October 2006

Fedora Extras

  • JasonTibbitts kindly approved my libglade package; he also spotted to non-weak-symbol issues, which I fixed

Tuesday 3rd October 2006

More Disk Failures

Woke to find that our new file and print server (my old desktop box) was having issues. One of the hard disks ib the RAID1 set had finally given up the ghost (there had been uncorrectable read errors from time to time for a while) and I was also getting some I/O errors on the other drive. Need to look at that when I get home from work.

Wiki

Fedora Extras

  • Rebuilt gtkwave, gtorrentviewer, and lat for rawhide due to there being a bug in the gcc version used to build the current releases; not sure that it was necessary to rebuild lat (a mono application) but the automated rebuild script won't be clever enough to know that either so I rebuilt it anyway

Local Packages

Wednesday 4th October 2006

Wiki

Wrote AdobeReaderOnFedora. Not something I particularly uee myself (I'm happy enough with Evince) but I'm sure other people will find it useful. I was asked about installing this on fedora-selinux-list and I couldn't find anything similar after a quick search.

Thursday 5th October 2006

Local Packages

  • Updated perl-Module-CoreList to 2.09

Monday 9th October 2006

Local Packages

  • Updated gtkwave to 3.0.13

  • Updated perl-Devel-Symdump to 2.0604

  • Updated perl-Test-Exception to 0.23

Fedora Extras

  • Updated gtkwave to 3.0.13

Tuesday 10th October 2006

Leon's First Nursery Visit

Leon went to spend an hour at nursery today to help him settle in. The nursery nurse took him straight into the room and we didn't say goodbye, which might have upset all parties. So we went off for a walk past the golf course to the Mersey Valley. When we got back, we were told he'd been alright for about twenty minutes and then seemd to notice we'd gone and started crying, which continued until we returned.

We'll have another go next Tuesday afternoon before his scheduled start date of next Thursday.

Local Packages

  • Updated dovecot to 1.0.rc8

  • Updated perl-Test-Exception to 0.24

Wednesday 11th October 2006

Local Packages

  • The BitTorrent download site has a new tarball, version 4.24.2, and RPMs of version 4.26.0. With no source available for the latter, I took a look at the former. There's quite a lot of changes, so I had to rework three of the patches I apply (to turn off the version check, to fix the application directory name, and to move the public key directory). The GUI seems to work OK on FC5, but the package is broken for python 2.3 now as it uses collections.deque, which is only available from python 2.4. The seed client doesn't work at all, falling over as follows:

    Traceback (most recent call last):
      File "/usr/bin/btseed", line 82, in ?
        config['torrent_dir'] = \
    AttributeError: 'module' object has no attribute 'decode_from_filesystem'
  • So I've decided not to release this version, and see if the sources for 4.26.0 turn up.

Thursday 12th October 2006

Local Packages

  • The source for bittorrent 4.26.0 is now available, which is good. It's no better than 4.24.2 though, falling over in much the same way. Here's another example, this time, the curses client:

    >>> unable to setrlimit  not allowed to raise maximum limit
    Traceback (most recent call last):
      File "/usr/bin/bittorrent-curses", line 45, in ?
        from BitTorrent.RawServer_twisted import RawServer, task
    ImportError: cannot import name task
  • One wonders what sort of QA process these so-called stable releases are going through.

Friday 13th October 2006

Friday 13th Unlucky for Leon

With Leon approaching being 13 months old, he's been struck by his first proper illness. I left work early to take him to see the doctor, who told us that Leon has an infection in his right ear. It's clearly causing him some distress and he's puking up most of his solids.

Saturday 14th October 2006

Leon's Illness

He's no better this morning so we used the prescription for amoxycillin that the doctor gave us yesterday. It has a very strong lemon smell and Leon won't have it. In fact he's so opposed to having it that he pukes it back up immediately when spoon-fed it. However, he has managed to keep it down when it's mixed with other stuff that has a taste that can mask it, such as Petits Filous fromage frais, or the Heinz Cauliflower and Broccoli Cheese.

Monday 16th October 2006

Local Packages

  • Updated dovecot to 1.0.rc9

Tuesday 17th October 2006

Local Packages

  • Updated perl-BerkeleyDB to 0.31

Wedbesday 18th October 2006

Local Packages

  • gtkwave updated to 3.0.14

  • perl-BerkeleyDB rebuilt as I forgot to sign yesterday's update

Fedora Extras

  • gtkwave updated to 3.0.14

Leon's Illness

The ear infection has now mostly cleared but he's still got a cold - cough and runny nose.

Thursday 19th October 2006

Fedora Extras

  • Updated perl-Math-Pari to 2.010708

Local Packages

  • Updated libidn to 0.6.8

Friday 20th October 2006

Fedora Extras

  • Updated lat to 1.2.1.1, after trying 1.2.1, raising upstream Bug #77 and getting a fix for it

Saturday 21st October 2006

Local Packages

  • Updated getmail to 4.6.5

Tuesday 24th October 2006

Fedora Extras

A tidy-up of the repositories revealed broken dependencies for smbldap-tools on the x86_64 architecture for Fedora Core 6 and development; this was due to one of the dependency packages, perl-Unicode-MapUTF8, failing to build on x86_64 and thus being excluded from that architecture. I have therefore had to convert smbldap-tools to an architecture-specific package and exclude it from being built for x86_64 as well.

Wednesday 25th October 2006

Local Packages

  • Updated perl-Unicode-Map8 to include the patches from openSUSE that fix it for the x86_64 architecture. Hopefully the version in Extras will do the same and then smbldap-tools can revert to being a noarch package.

Thursday 26th October 2006

Leon

Today was Leon's first full day at nursery. Things didn't look too good when he started crying as soon as the nursery nurse took him from me to take him into the play room, though he stopped crying soon afterwards. He also didn't get very involved in activities, nor did he eat much of anything all day (though he made up for that when he got home). However, when we arrived to collect him (when he'd been there around 8 hours), he was with the other babies in the room who were gathered in a circle playing together quite contently. When he saw us he burst into tears and crawled over to us without prompting, which was rather sweet. He soon cheered up though and didn't seem at all traumatised by the experience.

Fedora Upgrades

Updated both my home and work desktop boxes (x86_64 and i386 respectively) from Fedora Core 5 to Fedora Core 6. These have probably been the least painful upgrades I can ever remember doing. The work box was almost completely without issues and there were just a few niggles with the home machine upgrade:

  1. The named service didn't start up at boot time. Since this box uses a local caching nameserver for all host lookups, this broke lots of network services (samba mounts didn't happen for instance). Updating to the bind-* packages already in the updates repository fixed this.

  2. The squid cache wouldn't start, which was due to an incompatibility betwen the FC6 squid and FC5 config file. Merging my config file entries into the squid.conf.rpmnew file fixed this.

  3. Any attempt to run mock (my home desktop box is also my package build system) resulted in a kernel oops:

    Oct 26 15:47:39 metropolis kernel: Oops: 0000 [3] SMP 
    Oct 26 15:47:39 metropolis kernel: last sysfs file: /block/loop9/stat
    Oct 26 15:47:39 metropolis kernel: CPU 0 
    Oct 26 15:47:39 metropolis kernel: Modules linked in: tun cifs ipv6 autofs4 rfcomm l2cap sunrpc cpufreq_ondemand nls_utf8 loop video sbs i2c_ec button battery asus_acpi ac parport_pc lp parport floppy snd_intel8x0 snd_ac97_codec snd_ac97_bus snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device i2c_nforce2 snd_pcm_oss snd_mixer_oss ide_cd hci_usb ohci1394 i2c_core sg snd_pcm ieee1394 cdrom k8_edac edac_mc pcspkr snd_timer snd shpchp soundcore serio_raw forcedeth snd_page_alloc bluetooth dm_snapshot dm_zero dm_mirror dm_mod sata_sil24 sata_nv libata sd_mod scsi_mod raid1 ext3 jbd ehci_hcd ohci_hcd uhci_hcd 
    Oct 26 15:47:39 metropolis kernel: Pid: 10485, comm: mb Not tainted 2.6.18-1.2798.fc6 #1
    Oct 26 15:47:39 metropolis kernel: RIP: 0010:[<ffffffff80323f85>]  [<ffffffff80323f85>] context_struct_to_string+0x7b/0x18b
    Oct 26 15:47:39 metropolis kernel: RSP: 0018:ffff81003ed51ce8  EFLAGS: 00010296
    Oct 26 15:47:39 metropolis kernel: RAX: ffff81007e950000 RBX: ffff81003ed51dfc RCX: ffffffffffffffff
    Oct 26 15:47:39 metropolis kernel: RDX: 00000000ffffffff RSI: 0000000000000010 RDI: ffff81007e565c69
    Oct 26 15:47:39 metropolis kernel: RBP: 0000000000000000 R08: 00000000ffffffff R09: 0000000000000000
    Oct 26 15:47:39 metropolis kernel: R10: ffff81007e954000 R11: 0000000000000000 R12: ffff81003ed51d88
    Oct 26 15:47:39 metropolis kernel: R13: 0000000000000000 R14: ffffffffffffffff R15: ffff81003ed51dd0
    Oct 26 15:47:39 metropolis kernel: FS:  00002aaaaaac7db0(0000) GS:ffffffff80609000(0000) knlGS:00000000f7fe26c0
    Oct 26 15:47:39 metropolis kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
    Oct 26 15:47:39 metropolis kernel: CR2: ffff81087e94fff8 CR3: 000000005099e000 CR4: 00000000000006e0
    Oct 26 15:47:39 metropolis kernel: Process mb (pid: 10485, threadinfo ffff81003ed50000, task ffff8100425b3040)
    Oct 26 15:47:39 metropolis kernel: Stack:  000001bf0000024e ffff8100425b3040 ffff81007e54ec40 ffff81007b859bf8
    Oct 26 15:47:39 metropolis kernel:  ffff81003ed51db8 ffff81003ed51d88 ffff81007b859bc8 ffffffff803275cb
    Oct 26 15:47:39 metropolis kernel:  ffff81002584c420 ffff8100286632c8 0000000000000000 ffff81007b859bc8
    Oct 26 15:47:39 metropolis kernel: Call Trace:
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff803252bb>] security_compute_sid+0x243/0x340
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8031b2f3>] selinux_bprm_set_security+0xa9/0x183
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8023d5e0>] prepare_binprm+0xc1/0xed
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8023eb1a>] do_execve+0x11c/0x242
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff802537cc>] sys_execve+0x36/0x4c
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8025c3c7>] stub_execve+0x67/0xb0 
    Oct 26 15:47:39 metropolis kernel: DWARF2 unwinder stuck at stub_execve+0x67/0xb0
    Oct 26 15:47:39 metropolis kernel: Leftover inexact backtrace:
    Oct 26 15:47:39 metropolis kernel: 
    Oct 26 15:47:39 metropolis kernel: 
    Oct 26 15:47:39 metropolis kernel: Code: 48 8b 3c d0 44 88 e8 f2 ae 4c 89 e7 89 c8 f7 d0 8d 2c 30 89 
    Oct 26 15:47:39 metropolis kernel: RIP  [<ffffffff80323f85>] context_struct_to_string+0x7b/0x18b
    Oct 26 15:47:39 metropolis kernel:  RSP <ffff81003ed51ce8>
    Oct 26 15:47:39 metropolis kernel: CR2: ffff81087e94fff8
    Oct 26 15:47:39 metropolis kernel:  <3>BUG: sleeping function called from invalid context at kernel/rwsem.c:20
    Oct 26 15:47:39 metropolis kernel: in_atomic():0, irqs_disabled():1
    Oct 26 15:47:39 metropolis kernel:
    Oct 26 15:47:39 metropolis kernel: Call Trace:
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8026929b>] show_trace+0x34/0x47
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff802692c0>] dump_stack+0x12/0x17
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8029dc68>] down_read+0x15/0x23
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff80296254>] blocking_notifier_call_chain+0x13/0x36
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff80214e4b>] do_exit+0x1f/0x8c2
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff802648c0>] do_page_fault+0x794/0x80f
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8025ccdd>] error_exit+0x0/0x84
    Oct 26 15:47:39 metropolis kernel: DWARF2 unwinder stuck at error_exit+0x0/0x84
    Oct 26 15:47:39 metropolis kernel: Leftover inexact backtrace:
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff80323f85>] context_struct_to_string+0x7b/0x18b
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8032407c>] context_struct_to_string+0x172/0x18b
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff803275cb>] mls_compute_sid+0xe6/0x1c6
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff803252bb>] security_compute_sid+0x243/0x340
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8031b2f3>] selinux_bprm_set_security+0xa9/0x183
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8023d5e0>] prepare_binprm+0xc1/0xed
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8023eb1a>] do_execve+0x11c/0x242
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff802537cc>] sys_execve+0x36/0x4c
    Oct 26 15:47:39 metropolis kernel:  [<ffffffff8025c3c7>] stub_execve+0x67/0xb0
    Oct 26 15:47:39 metropolis kernel:
    Oct 26 15:48:39 metropolis kernel: security:  3 users, 6 roles, 1562 types, 166 bools, 1 sens, 1024 cats
    Oct 26 15:48:39 metropolis kernel: security:  58 classes, 48237 rules
    Oct 26 15:48:39 metropolis kernel: NMI Watchdog detected LOCKUP on CPU 1
    Oct 26 15:48:39 metropolis kernel: CPU 1
    Oct 26 15:48:39 metropolis kernel: Modules linked in: tun cifs ipv6 autofs4 rfcomm l2cap sunrpc cpufreq_ondemand nls_utf8 loop video sbs i2c_ec button battery asus_acpi ac parport_pc lp parport floppy snd_intel8x0 snd_ac97_codec snd_ac97_bus snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device i2c_nforce2 snd_pcm_oss snd_mixer_oss ide_cd hci_usb ohci1394 i2c_core sg snd_pcm ieee1394 cdrom k8_edac edac_mc pcspkr snd_timer snd shpchp soundcore serio_raw forcedeth snd_page_alloc bluetooth dm_snapshot dm_zero dm_mirror dm_mod sata_sil24 sata_nv libata sd_mod scsi_mod raid1 ext3 jbd ehci_hcd ohci_hcd uhci_hcd
    Oct 26 15:48:39 metropolis kernel: Pid: 10607, comm: load_policy Not tainted 2.6.18-1.2798.fc6 #1
    Oct 26 15:48:39 metropolis kernel: RIP: 0010:[<ffffffff8025fd0f>]  [<ffffffff8025fd0f>] __write_lock_failed+0x7/0x20
    Oct 26 15:48:39 metropolis kernel: RSP: 0018:ffff81001118faf0  EFLAGS: 00000083
    Oct 26 15:48:39 metropolis kernel: RAX: 0000000000000001 RBX: ffffffff8056e5a0 RCX: 0000000000000000
    Oct 26 15:48:39 metropolis kernel: RDX: 00000000ffffffff RSI: 0000000000000246 RDI: ffffffff8056e5a0
    Oct 26 15:48:39 metropolis kernel: RBP: ffff81001118fde8 R08: ffff81003d005f00 R09: ffff81001118fde8
    Oct 26 15:48:39 metropolis kernel: R10: 0000000000000000 R11: 0000000000000001 R12: ffff81001118fe18
    Oct 26 15:48:39 metropolis kernel: R13: ffff81001118fc78 R14: 0000000000000015 R15: 0000000000000001
    Oct 26 15:48:39 metropolis kernel: FS:  00002aaaaaac8380(0000) GS:ffff81007fea89c0(0000) knlGS:00000000f7fdb6c0
    Oct 26 15:48:39 metropolis kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
    Oct 26 15:48:39 metropolis kernel: CR2: 00000034e98178f0 CR3: 000000000da63000 CR4: 00000000000006e0
    Oct 26 15:48:39 metropolis kernel: Process load_policy (pid: 10607, threadinfo ffff81001118e000, task ffff810055981080)
    Oct 26 15:48:39 metropolis kernel: Stack:  ffffffff8022e8e6 00000000fffffff4 ffffffff80324fb9 ffff8100437a38c0
    Oct 26 15:48:39 metropolis kernel:  0000000000000003 ffff810040df92e0 000000000000003a ffff81004f31a8c0
    Oct 26 15:48:39 metropolis kernel:  0000000000000006 ffff8100057c5800 000000000000061a ffff8100057c5ba0
    Oct 26 15:48:39 metropolis kernel: Call Trace:
    Oct 26 15:48:39 metropolis kernel: Inexact backtrace:
    Oct 26 15:48:39 metropolis kernel:  [<ffffffff8022e8e6>] _raw_write_lock+0x69/0x83
    Oct 26 15:48:39 metropolis kernel:  [<ffffffff80324fb9>] security_load_policy+0x1d6/0x285
    Oct 26 15:48:39 metropolis kernel:  [<ffffffff8031e7b8>] sel_write_load+0xbc/0x31b
    Oct 26 15:48:39 metropolis kernel:  [<ffffffff802162b6>] vfs_write+0xce/0x174
    Oct 26 15:48:39 metropolis kernel:  [<ffffffff80216b26>] sys_write+0x45/0x6e
    Oct 26 15:48:39 metropolis kernel:  [<ffffffff8025c181>] tracesys+0xd1/0xdc
    Oct 26 15:48:39 metropolis kernel: 
    Oct 26 15:48:39 metropolis kernel:  
    Oct 26 15:48:39 metropolis kernel: Code: f3 90 81 3f 00 00 00 01 75 f6 f0 81 2f 00 00 00 01 0f 85 e2
    Oct 26 15:48:39 metropolis kernel: console shuts up ...
  4. The presence of "selinux" and "security" near the top of the stack trace led me to suspect SELinux as being a possible culprit, and this turned out to be correct. I rebuilt my local policy module packages (the .pp files) and re-added them to policy:

    # semodule -i localmisc.pp -i mock.pp
  5. This fixed the problem - it would seem that FC5 policy module packages don't mix well with FC6 base policy.

The work machine had the same mock issue (and was fixed the same way), plus I needed to download and install the vmware-any-any-update104.tar.gz update to get VMware Server working.

Friday 27th October 2006

Fedora Extras

  • Fixed the scriptlets for ORBit-devel to ensure that they didn't cause a failure exit status (and hence failed transactions resulting in multiple packages being recorded in the RPM database)

  • Updated perl-Math-Pari to 2.010709, but only in development; there's no changes in this release that make pushing an update for current distributions worthwhile

Local Packages

Updated my buildsystem so that I can build "pukka" Fedora Core 6 packages and development packages are now tagged .fc7. The main non-triviality here was that I had to pull yum-arch from a Fedora Core 5 box so as to be able to create the metadata for legacy distributions, since yum-arch is no longer available in Fedora Core 6. I'll eventually get round to packaging yum-arch up as a separate package and maybe even submit it to Extras.

Sunday 29th October 2006

Fedora Extras

  • perl(Unicode::MapUTF8) is now available on x86_64 so I removed the ExcludeArch: x86_64 from smbldap-tools and rebuilt it as an architecture-independent package, and closed Bug #211866

Monday 30th October 2006

Fedora Extras

  • Updated gtkwave to 3.0.15

Local Packages

  • Updated gtkwave to 3.0.15

  • Updated curl to 7.16.0; this release of curl bumps the soname of libcurl to libcurl.so.4, which would cause dependency problems when installed on all current and legacy distributions, so I also introduced a new compatibility package, libcurl7155, to provide libcurl.so.3 to satisfy the dependencies

Tuesday 31st October 2006

Local Packages

  • Updated moin to 1.5.6

  • Updated mod_fcgid to 2.0

  • Updated city-fan.org-release to include a separate city-fan.org-development.repo file for the development repo; the regular repo is enabled by default and the development repo is disabled by default, except in the development repo itself, where it's the other way around (i.e. just like in the fedora-release package)

  • Raised Bug #213260 regarding FC6's yum sometimes installing the wrong packages when trying to resolve dependencies

Wiki

  • Updated the wiki to MoinMoin version 1.5.6, running on mod_fcgid 2.0

Fedora Extras

  • Rebuilt gtorrentviewer for Rawhide so as to pick up the new libcurl.so.4 dependency

  • Updated my python-zope-interface submission to include running the test suite in %check, and the submission was subsequently approved by JeffOllie

  • Updated mod_fcgid to 2.0, but only for Rawhide; will add to the older branches in a week or so if there are no issues with it

Previous Month: September 2006
Next Month: November 2006

Recent