Index Thing


The Hippo Techo Blog


112 Posts. Displaying page 12 of 12.

Go to page:

Hide All

Software - OSX
May 08, 2016

Mac Mini 2012 – Upgrade from 10.9.5 (Mavericks) to 10.11.3 (El Capitan).

Close ...
Decided to uprade the mini since Mavericks only supported up to Python 2.7 and I needed Python 3. Rather than try to shoehorn Python 3 into Mavericks an upgrade to El Capitan was done (seems to be reasonably OK as an OS).
An upgrade was done rather than a clean install and appears to have gone quite well. All of the connected devices work OK and no real showstoppers yet.


Aperture 3.5.1 broke – needed to upgrade to 3.6
iPhoto broke – update allowed when El Capitan was installed.
Note: both these apps are of course legacy with the new Photos app being used.

iTunes – this was the biggest hassle as I had uninstalled iBooks on Mavericks to allow iTunes to manage my books/pdfs but of course the El Capitan upgrade re-installs iBooks and removes the titles from iTunes.
I tried to customise iTunes again but El Capitan has SIP technology which hinders the removal of system related files/services so I have to bite the bullet and use iBooks. Some changes have been done to iBooks to allow metadata changes but it is still a very primitive app compared to iTunes which managed all my media very satisfactorily. It almost seems that Apple is more intent on keeping the ‘walled garden’ intact than improving the usability of their media apps.

Ref: te/

The 10.11.4 update just released installed OK.

Application – PDFSam broke – needed to install Java 6 runtime

Application – RSSOwl broke – needed to install Java 6 runtime

Xcode – for some reason 7.2.1 needed to update twice

Application – LaunchPadManager – app pages in app don’t equate with LaunchPad pages
Eg: if 6 pages of apps in LaunchPad then use pages 2 ~ 7 in LaunchPadManager and result in layout and LaunchPad will be pages 1 ~ 6

Ethernet 0 was nonoperable after mini had been in sleep mode (doesn’t happen every time).
Network Preferences showed the connection status as Connected and currently active with an IP address but not working – both DNS and pinging was non-functional. Under the WINS parameters the Workgroup was showing as WORKGROUP (default) and although it could be changed the change would not be applied.

Using Terminal –

sudo ifconfig en0 down

sudo ifconfig en0 up

Ethernet interface working again and workgroup back to the previous name (HOMELAN).

All in all not too painful but I will be trying hard to live with the crappy iBooks and watching for the ethernet problem – oh and yes the ‘look’ of El Capitan is very much ‘it grows on you’ compared to the more polished look of Mavericks (some of the icons in the dock look rather insipid).

Hardware - Mac
May 08, 2016

Disk Failure with LaCie 2Big Thunderbolt RAID

Close ...
Had a disk go bad in the LaCie 2Big and proved interesting. The short story is that I started with a 2TB RAID 1 arrangement and now have two standalone 2TB drives in a RAID box.
Why so? There was so much messing about with the RAID when a disk failed that it is easier to use two disks with a manual mirror process.

The LaCie 2Big is a 4TB unit using RAID 1 (Disk Mirror) giving a mirrored 2TB arrangement – all well and good when the disks are OK. The controller actually uses Apple RAID within OSX which I discovered on my internet trawl is no longer supported in El Capitan – a recommended replacement is SoftRAID Lite/Pro. At this stage the SoftRAID product will not convert an Apple RAID set (they are working on it but the raid set will need to be working correctly) so a fresh install is recommended which means fully backing up the RAID which in my case was about 1 TB worth of data.
So back to the issue – one disk had degraded and a raid slice gone missing. Because I had issues with the LaCie Raid Monitor being installed on Mavericks it was uninstalled so the only way of diagnosing the disks was to use the OSX Disk Utility. The problem with Apple Raid is that it doesn’t show clearly which disk is the degraded one althought DU does show an amber colour for the degraded disk. I checked all the fine print then went for the downpower and pull the disk option – like any binary choice I always pick the wrong one so downpower again and pull the faulty one!
As an indicator the disks originally showed as disk 4 and disk 6 and as it transpired disk 6 was D1 in the RAID box and disk 4 was D2 – disk 6 was the bad one. At this point because I hadn’t enabled Automatic Raid Rebuild as an option in the raid set a manual rebuild was required. It seems that Disk Utility is OK at creating new raid sets but not so good at rebuilding – using the GUI it is not clear on the exact procedure. I think it can be done via the terminal successfully but I persisted with the GUI. Trying to delete the missing slice I managed to delete the raid set which then leaves a copy of the data on each disk as a standalone copy. The new disk was inserted and partitioned as per the good disk and at this point I decided to keep each disk as a standalone drive in the LaCie box and use a synchronisation app to manually mirror the disks.
With hindsight if I was to use RAID again I would go with SoftRAID (AppleRAID is defunct) and hopefully the change disk/auto rebuild/mirror would work fine however I need storage to work well so using multiple standalone drives seems the most reliable option at this point.

Note: The LaCie RAID Monitor app has been updated so maybe this would also fix the issues with Mavericks and later.


I’ve since upgraded to El Capitan (10.11.4) but have decided to stick with just using the LaCie RAID as two separate disks. However the options for using RAID on El Capitan are –

AppleRAID is discontinued but Internet Recovery mode still allows use of the /R option on reboot which gives access to the old AppleRAID tools

Use of a procedure via the Command Line Interface in Terminal

Use a third-party product called softRAID which is the method I would probably go for if I did RAID again( they have a cheaper product available for El Capitan users).

Some good refs are – n_US&wwwlocale=en-us

May 08, 2016

BT Home Hub 4 as an Access Point on EE BrightBox Router

Close ...
These days using EE Broadband and unfortunately with the 16M service the Brightbox 2 router (with 1 Gbps port) is not available and the Brightbox router only has 4 x Fast Ethernet ports. This means the 1 Gbps core LAN network is degraded to Fast Ethernet (100Mbps) however it is possible to use the HH as an access point.

HH4 –
Disable DHCP
Allocate IP outside DHCP range

eg: EE router/BB modem –
HH4 router/hub -
Both routers linked via Fast Ethernet (100 Mbps)
HH4 Gigabit port linked to Gigaswitch/homelan

Clients -
IP: 192.168.1.x or DHCP (will use DHCP from EE router)
Sub Mask:

This will also allow a Gbps connection from the LAN to the HH4 giving a Giga core at layer 2 but the Fast Ethernet link between routers may still affect the core on traffic routed via the EE router.

Downsides –
Red BB light on HH4 will flash because there is no connection to the BB
Firmware in HH4 will not upgrade as it needs to be connected to BT BB to do so.

Software - OSX
Nov 18, 2015

OSX - Open With Context in Finder Sluggish with Beachball

Close ...
Using Mavericks (10.9.5) I found when hovering over the Open With in the Finder Right-Click context that a beachball/pizza wheel would appear for quite some time before the App Menu popped up. The Finder seemed to be waiting for the external RAID drives to spin up.

Fix: A reset of the Launch Services database was required –

/System/Library/Frameworks/CoreServices.framework/Versions/A/Fram eworks/LaunchServices.framework/Versions/A/Support/lsregister -kill -r -domain local -domain system -domain user

Enter the above into Terminal and bingo – instant App Menu popups again!

Further detail see -

Software - OSX
Nov 18, 2015

OSX - VirtualBox Support for Windows Guests on OSX Hosts

Close ...
VirtualBox 5.0 introduced bidirectional Drag and Drop but unfortunately only worked in the Host to Guest direction. VB 5.0.2 fixed the issue but broke the Host to Guest direction causing the VM to abort when a DnD was done.

Update: DnD in the Host to Guest direction fixed in 5.0.4 load.

Software - OSX
Nov 18, 2015

OSX - Geotagging .dng Files Broken

Close ...
My cameras do not have inbuilt geotagging capability however the longitude/latitude details can be added later.

Apps suitable for OSX –

GeoTag for OSX -
geotag - java based app

These apps allow the addition of geotagging data in post-process by placing a pin on the map however to work for DNG the camera model must be supported by the OSX. eg: a DNG from a Leica XVario will require Yosemite but a Leica X2 DNG will work on Mavericks. JPG’s are fine and I guess other RAW formats are in a similar situation as requiring OSX level support.

Software - OSX
Nov 18, 2015

Firefox Nightly - Web Page Complete Save is Broken

Close ...
Page saves will not complete and show 0% in download box. Option to cancel download will be given on closing Nightly. This appears to have happened on later loads of FF. Web Page Complete is not given as an option, only HTML Page.

Fix: If E10S multi-processing is disabled then Web Page Complete functionality will again be available.

Note: Web Save Complete is provided by the Mozilla Archive Format add-in

May 19, 2015

FF/Nightly – JS scroll broken – get TypeError: Value can’t be converted to a dictionary on hover.

Close ...
Noticed the other day that the JS scrolling on the hippo was not working with FF or Nightly.

FF Dev Edition 39 and Nightly 41 both had the problem – FF 24 was OK.
It appears that the newer browsers objected to the scroll(1) function and once it was renamed everything worked OK.

<a href=”#” onmouseover=”scroll(1);” onmouseout=”noScroll();” onclick=”return false”>UP<span> … </span></a>

<a href=”#” onmouseover=”goScroll(1);” onmouseout=”noScroll(0);” onclick=”return false”>UP<span> … </span></a>


Found later on that the new code was breaking on Safari 8.0.6 on Yosemite but was OK with Safari 7.1.6 on Mavericks. The error produced was –

ReferenceError: Can’t find variable: goScroll

Interestingly the later Safari worked with the old code.

The final fix was to upgrade the code to use listeners and it works fine.

<a href=”#” id=”navup” onclick=”return false”>UP<span> … </span></a>
Listener –
element = document.getElementById(‘navup’);
element.addEventListener(“mouseover”,function(){ scroll(1); });

Software - OSX
May 19, 2015

OSX – Accessing iCloud Calendar on Snow Leopard (10.6.8)

Close ...
iCloud support really started on Lion (10.7) but a workround can be done to allow Snow Leopard to work.

Need to obtain PXX number from calendar URL:
Use browser and log in to and open Calendars
Click the Share symbol next to desired calendar > click Public > click Share
The shared calendar's URL will display and include a PXX number – once you know this the calendar can be unshared if desired.

Setup iCal account:
Open iCal > Preferences > Accounts
Create a new account -
Type: CalDav
User: Apple ID (iCloud username)
Port: 443
SSL: Checked

Should work but check that iCal is Online working – if so the calendar should appear and update in the sidebar under the Apple ID name.

403 errors may occur once logged in and this may be due to the server URL being incorrect (eg: wrong PXX number).
Selecting Manual updates and going Offline will suffice until the server details are corrected.

Software - OSX
May 19, 2015

WireShark crashes on start with Yosemite (10.10.2)

Close ...
WireShark would not run and would close down immediately.

Yosemite breaks the symbolic link for the folder paths. Since Apple no longer produces the X11 app users now have to install XQuartz X11 and up till recently the path was /usr/X11 – with XQuartz it is now /opt/X11 and /usr/X11 is just a symbolic link to /opt/X11. Yosemite does not consider this so I had to re-install XQuartz X11 and WireShark was then fine.

112 Posts. Displaying page 12 of 12.

Go to page: