FUJ00081071
FUJ00081071
D'Alvarez
Cochrane
in Achte
“)); Turner lan
Salawu Saheedf. “} Lywood Patt”
Hammond ChristopherI
Porter Stevenf”” ‘
Welsh Graham[/O=EXCHANGE/OU=ADMINGROUP1/CN=RECIPIENTS/CN=WELSHG]
4 Wed 6/16/2010 7:24:38 PM (UTC)
Subject: I RE: Rates Board & Region
HNGx Region Risk.xIsx
HNGx Region Risk.xIs
Okay based on conversations with Graham & Mark and the information that I have available to me, Ref Data Enablement and Non Region 1 data,
I have produced the model below.
This is a worst case and can be fine tuned once the actual Event and frequency are known.
Counter Trading Days for Non Region 1 Branches
Total Counter A
05Jun O06Jun O7Jun = 08Jun = 09Jun = 10-Jun Trading Days Bor
Migrated Counters Volumes 177 16 1177 1177 4177 A177
for Week 5th - 10th June 5901
Wc 7th June O7Jun =O08Jun =09Jun =10Jun 11Jun = 12Jun = 13-Jun
Previous Migrations AV77 1177 1455 1455 1855 1855 16 8990
Migration Activity 278 400
Ave Events per Day
Wc 14th June 14Jun 15Jun 16Jun = 17Jun = 18Jun = 19-Jun = 20-Jun
Previous Migrations 1855 1855 2289 2517 2966 2966 16 14464
Migration Acti 434 228 449
Ave Events per Day
Wic 21 June 21Jun 22Jun 23Jun = 24un-25Jun = 26-Jun = 27-Jun
Previous Migrations 2966 3193 3504 3504 3805 3605 47 20624
Migration Activity 227 M1 301
Ave Events per Day
Note: Figures in Blue are projected worse case based on the information available
In addition PSA excel file if which has the background data:
in older format
Can we have the “Event” identified so that we can ensure that the SMC are proactively aware what they are looking for.
I appreciate and recognise that any attempt to pull forward will impact more than just R1 activities and thus we need to validate and contain as
much as we can.
Regards,
Migration Governance Manager
Royal Mail Group Account
FUJITSU
FUJ00081071
FUJ00081071
Lovelace Road, Bracknell, Berkshire RG12 8SN
Mob: [
E-mail:
Web
ik Fat
st Please consider the environment - do you really need to print this email?
Fujitsu Services Limited, Registered in England no 96056, Registered Office: 22 Baker Street, London, W1U 3BW
This e-mail is only for the use of its intended recipient. Its contents are subject to a duty of confidence and may be privileged. Fujitsu does not guarantee that
this e-mail has not been intercepted and amended or that it is virus-free.
From: Allen Graham (BRAO1)
Sent: Wednesday, June 16, 2010 4:01 PM
To: Butts Geoff; Welsh Graham; D'Alvarez Alan; Andrews Mark (GD HQ); Cochrane Vince; Davidson James; Van Achte Gaetan;
Richardson Debbie DB; Turner Ian T
Cc: Salawu Saheed; Lywood Pat; Bamber Sheila; Hammond Christopher; Jepson Mark; Porter Steven
Subject: RE: Rates Board & Region
Development have advised that hot fixing 1.23.00 will be painful
To avoid issues of creating regression we would really have to deliver the hotfix for 1.23.00 before the hotfix for 1.23.01. This will
therefore impact the date of the latter fix and likely cause a delay to LST for 1.23.01 and other important fixes and also delay the
relaase of the fix into the R2 baseline. In short we are doing so much that something would have to give to do this. If at all possible
we should therefore avoid this.
After discussion there is an event generated which shows a counter has reverted to boot strap data; and form looking at events we
have form the last few days we have calculated that approx 0.004% of counters not in region 1 will get this error daily.
So from Gareth...
I would suggest the following approach:
1. On day of Migration, any Branch not in Region 1 needs to reboot all counters. This needs to be in place through the entire
rollout or until we get a fix in the migration software target.
2. Until we have 1.23.01.01 rolled out to all counter, then need to get SMC to monitor for the Event that indicates reversion to
bootstrap. They should then check these against non Region 1 Branches (RDT can provide as a spreadsheet) and if
necessary phone the branch asking them to reboot that counter that generated the event.
This seems a reasonably watertight approach and covers all the known problems in this area.
If we follow this approach we need to define the exact event to look far and provide the Bureau region info to SMC and they may
well need to change the event filters to stop filtering out the event that is generated.
Regards,
Graham Allen
Application Services - Post Office Account
FUJITSU
E-mail
Web: htfp://uk
wa Please consider the environment - do you really need to print this email?
Fujitsu Services Limited, Registered in England no 96056, Registered Office: 22 Baker Street, London, WiU 38W
FUJ00081071
FUJ00081071
This e-mail is only for the use of its intended recipient. Its contents are subject to a duty of confidence and may be privileged. Fujitsu Services does not
guarantee that this e-mail has not been intercepted and amended or that it is virus-free.
From: Butts Geoff
Sent: 16 June 2010 14:39
To: Allen Graham (BRAO1); Welsh Graham; D'Alvarez Alan; Andrews Mark (GD HQ); Cochrane Vince; Davidson James; Van Achte
Gaetan; Richardson Debbie DB; Turner Ian T
Cc: Salawu Saheed; Lywood Pat; Bamber Sheila; Hammond Christopher; Jepson Mark
Subject: RE: Rates Board & Region
Importance: High
Graham,
Thanks. As per our discussion I would like Graham W and Mark to agree:-
1. Is there a viable workaround for branches (i.e. rebooting the counter immediately after migration)?
2. If we go with targeting the fix into 1.23.01.01, this will take approximately 4 weeks to get into live (we'll be in the 3° week of
rollout). Will this be acceptable from a support and business perspective?
We'll review this on Morning Prayers tomorrow before any communication is made to POL.
Regards,
Geoff
From: Allen Graham (BRAO1)
Sent: 16 June 2010 13:36
To: Welsh Graham; D'Alvarez Alan; Andrews Mark (GD HQ); Cochrane Vince; Davidson James; Van Achte Gaetan; Butts Geoff;
Richardson Debbie DB; Turner Ian T
Cc: Salawu Saheed; Lywood Pat; Bamber Sheila; Hammond Christopher; Jepson Mark
Subject: RE: Rates Board & Region
Importance: High
Graham,
We need to be clear then what we are proposing.
So what I understand is:
1. The first day of trading on HNGx these branches WILL have the wrong rates
2. There is a small risk that at any other time single counters could fallback to boot strap data invisibly and have the wrong rates
To protect against 2 earlier we will need to develop a hot fix against 1.23.00 (1.23.00.01?) This is the release that is currently in or
about to go to Model Office and would seem the earliest sensible delivery release. I have asked development for a date for this not
affecting the agreed plan for 1.23.01.01 already in track. LST will need to confirm how they test this and how long.
Additionally to protect against 1 we will need to update the migration package. This will take a couple of days to re-engineer the
migration package to contain 1.23.01.01 and will require LST to do redo the migration testing which they need to confirm but I believe
is 2 weeks.
Finally I believe we identified a problem that we could not simply cut over from one migration package to a new one whilst migrations
were taking place and hence this may require a pause in rollout. This needs to be confirmed but not sure by who.
How do we agree this and who manages it?
Regards,
Graham Allen
Application Services - Post Office Account
FUJ00081071
FUJ00081071
FUJITSU
Lovelace Road, Bracknell, Berkshire, RG12 8SN.
Tel: internally {GRO
Mob:
E-mai
Web: ik fujits
st Please consider the environment - do you really need to print this email?
Fujitsu Services Limited, Registered in England no 96056, Registered Office: 22 Baker Street, London, W1U 3BW.
This e-mail is only for the use of its intended recipient. Its contents are subject to a duty of confidence and may be privileged. Fujitsu Services does not
guarantee that this e-mail has not been intercepted and amended or that it is virus-free.
From: Welsh Graham
Sent: 16 June 2010 12:12
To: Welsh Graham; D'Alvarez Alan; Allen Graham (BRAO1); Andrews Mark (GD HQ); Cochrane Vince; Davidson James; Van Achte
Gaetan; Butts Geoff; Richardson Debbie DB; Turner Ian T
Cc: Salawu Saheed; Lywood Pat; Bamber Sheila; Hammond Christopher; Jepson Mark
Subject: RE: Rates Board & Region
All,
To be clear I have spoken to Saheed & agreed the following:
¢ HSD & SMC need to be aware of the 119 this week.
e In conjunction with Ref data establish the numbers of Branches that could be impacted for next week.
Spoken to Mark Andrews:
e Agreed that nothing should be said to POL until we are clear on what can be achieved above and beyond the targeting of Ctr 1.23.01,
agreed this morning.
e Briefed him on the detail above.
Spoken to lan and left a message for Graham A:
e Requesting a view on what could be achieved in pulling forward a fix prior to 1.23.01
Texted Geoff:
¢ He is aware of the headline detail
¢ However as he and Alan’s mailboxes are full not sure if he has seen the mails etc.
From: Welsh Graham
Sent: Wednesday, June 16, 2010 11:36 AM
To: Alan D'Alvarez; Allen Graham (BRAQ1) ; Andrews Mark; Cochrane Vince; Davidson James
i Gaetan Van Achte; Geoff Butts; Richardson Debbie DB
Tan T
Cc: Salawu Saheed
Subject: Rates Board & Region
Turner
All,
FUJ00081071
FUJ00081071
Further to the dialogue on this last night and going through my mail I had reason to question the 119 quoted given the source and title of the
data set.
Whilst I appreciate that we have taken steps to target the fix the issue is that it is not 119 Branches that could be effected BUT 6258, including
CTO’s.
As such we need to revisit the approach to fix urgently.
Regards,
From: Woodley Adam
Sent: Wednesday, June 16, 2010 10:09 AM
To: Welsh Graham; Parker Steve (PostOfficeAccount)
Cc: Young Nicola; Gelder Robert; Wilcox David; Bansal Steve (BRAO1); Salawu Saheed; Hill Sarah(POA HSD)
Subject: RE: region issues
Good Morning,
Please find attached a report detailing all offices which are not region 1, as requested.
Regards << File: ALL_OFFICES_NOT_REGION1.zip >>
Adam
From: Welsh Graham
Sent: 16 June 2010 09:06
To: Woodley Adam; Parker Steve (PostOfficeAccount)
Cc: Young Nicola; Gelder Robert; Wilcox David; Bansal Steve (BRA01); Salawu Saheed; Hill Sarah(POA HSD)
Subject: RE: region issues
Adam,
If you have not already can you please provide a view on ALL Branches i.e. the full Estate that are not Region 1
Or Confirm that the detail within the original report is all the Branches.
From: Woodley Adam
Sent: Tuesday, June 15, 2010 11:50 AM
To: Parker Steve (PostOfficeAccount)
Cc: Young Nicola; Welsh Graham; Gelder Robert; Wilcox David
Subject: FW: region issues
Good Morning,
Please find attached the HNG-x enablement monitor with the BDC Regions added, as requested.
Regards << File: HNGx_Enablement_Regions.xls >>
Adam
FUJ00081071
FUJ00081071
From: Parker Steve (PostOfficeAccount)
Sent: 15 June 2010 09:01
To: Gelder Robert; Wilcox David
Subject: RE: region issues
Can you supply a list please
From: Gelder Robert
Sent: 15 June 2010 09:00
To: Parker Steve (PostOfficeAccount); Wilcox David
Subject: RE: region issues
Good Morning,
I understand that all those currently migrated, have now resolved themselves. Regarding future migrations, we are able to
identify which branches are in the same scenario (i.e. which migrating branches are not in Region 1).
Regards
Rob
From: Parker Steve (PostOfficeAccount)
Sent: 15 June 2010 08:35
To: Wilcox David; Gelder Robert
Subject: RE: region issues
Gents,
Is it possible to identify the other sites that this problem will impact, i.e. a list of sites on this special bureau rate?
Steve
From: Wilcox David
Sent: 14 June 2010 17:27
To: Parker Steve (PostOfficeAccount)
Cc: Gelder Robert
Subject: FW: region issues
Steve,
Seems as though the problem has been diagnosed.
David
David Wilcox
Reference Data Manager, Retail and Royal Mail Group
Fujitsu
Fujitsu Services Limited, Registered in England no 96056, Registered Office 22 Baker Street, London, W1U 3BW
This e-mail is only for the use of its intended recipient. Its contents are subject to a duty of confidence and may be privileged. Fujitsu does not
guarantee that this e-mail has not been intercepted and amended or that it is virus-free.
FUJ00081071
FUJ00081071
From: — Macdonald Duncan
Sent: 14 June 2010 16:59
To: Wilcox David; McKeown Kevin
Subject: FW: region issues
From: Porter Steven
Sent: 14 June 2010 16:56
To: Macdonald Duncan
Cc: Mulye Vinay
Subject: region issues
Duncan,
We think we spotted a Counter code problem:
If we receive downloads in the order below (which we have seen previously, that we think we do), on a bootstrap counter,
freshly migrated:
Package 2
Package 1
Package 3
Package 4
Then the region update in package 1 is ignored, and when package 4 is received (the first time we go tradable) then the rates
board is updated with the old region, not the new region.
This issue goes away once counter is rebooted. The root cause seems to be that the region id is cached at startup, and the
logic to determine when to re-read it is wrong - current code only re-reads the region when the counter is already fully tradable,
and a new package’ is received.
We'd propose to reload the cached region whenever we receive a package’ update regardless.
Can you get the peak routed our way?
Regards
Steve
Steven Porter
Software Design & Development (SDD) Practice & HNGX Design/Development
Fujitsu Services
Berkshire, RG12 8SN