Mar 7, 2018 | Uncategorized
As part of the upgrades and improvements to our network we are moving our connectivity in this data centre to a new rack with dual power feeds and additional cooling.
We already have a new router and switch in this rack – However subject to resources we will be moving some of our hardware and links between the two racks this coming Sunday. The work will start in the late afternoon and into the evening.
During this time various parts of our network and services connected directly to this network PoP may experience a period of downtime. Traffic may also take alternative routes to some locations and there’s a chance of some congestion for some traffic.
This PoP currently hosts
– some Transit and peering links which will simply take alternative routes
– One of our DNS servers (the other servers will work fine)
– Some TalkTalk based ADSL traffic – in these cases PPP sessions will drop BUT should auto re-connect on the Telehouse link
– Approx 21 leased line circuits that will be contacted directly
We will confirm the date once we have confirmation of resources from the data centre provider to help move some fibre cross conencts
We will keep the downtime to a minimum and we will update this post as much as we can
Any questions please email support in the normal way
[update]
This work is completed. All services in HEX are now in our new rack and core network
Feb 15, 2018 | Information, Planned Work
Our supplier has identified a potential issue with a fibre bearer which holds some backup links to our Huntingdon Data Centre. They are making a config change to correct this overnight.
As these are backup links it should not have a significant impact but there is some chance of packet loss as the link drops / returns as the network re-converges.
We will update this post as the work progresses
Feb 9, 2018 | Planned Work
We have noticed a small issue on a switch port connected to some leased lines. The issue is not currently service affecting – and needs a simple one line config change to resolve.
As there is a very small risk to a few seconds of packet loss during the change this will be performed tonight
Only leased lines connected to a single bearer in Telehouse North will be affected and we don’t expect more than a few seconds of packet loss during the change
Any questions or issues please contact support via the normal means
Feb 9, 2018 | Information
Good morning.
Whilst we don’t supply Netgear routers, we know that a number of our customers use them, so wanted to ensure you were aware of potential vulnerabilities that need to be addressed via a firmware update.
The list of (so far) affected models is below. You or your IT specialist should ensure that you update as soon as possible to mitigate any risks.
List of NETGEAR Routers At Risk
DGN2200v4
R6100
R6220
R6250
R6300v2
R6400
R6400v2
R6700
R6900
R6900P
R7000
R7000P
R7100LG
R7300DST
R7500
R7500v2
R7800
D7800
R7900
R8000
R8300
R8500
D8500
WNDR3400v3
WNDR4500v2
EX6200v2
The Five Vulnerabilities
TWSL2018-002: Password Recovery and File Access
Trustwave SpiderLabs Advisory
NETGEAR advisory
Some routers allow arbitrary file reading from the device provided that the path to file is known. Total of 17 products are affected.
TWSL2018-003: Finding 1: Post-Authentication Command Injection
Trustwave SpiderLabs Advisory
NETGEAR advisory
This one affects six products and reflects a root level OS command execution via the device_name parameter on the lan.cgi page, although the attack requires authentication.
TWSL2018-003: Finding 2: Authentication Bypass
NETGEAR advisory
This also affects large set of products (17) and is trivial to exploit. Authentication is bypassed if “&genie=1” is found within the query string.
TWSL2018-003: Chained Attack: Command Injection
NETGEAR advisory
This is a three-stage attack leveraging three separate issues: CSRF token recovery vulnerability and the two findings in TWSL2018-003. As a result, any user connected to the router can run OS commands as root on the device without providing any credentials.
TWSL2018-004: Command Injection Vulnerability on D7000, EX6200v2 and Some Routers
Trustwave SpiderLabs Advisory
NETGEAR advisory
Only 6 products are affected, this allows to run OS commands as root during short time window when WPS is activated.
Feb 1, 2018 | Information, Planned Work
Following the rapid growth on the Merula network we have been planning a major upgrade to our core switching platform to Juniper devices. These will allow us to dramatically increase the capacity between sites and add further stability to our network. Over the coming week, work will take place to introduce them into our network.
This will in due course lead to some limited downtime as links are migrated. We will update here with specific slots for this but wanted to provide some initial details for this work.
Any initial questions, please email support via the normal means.
Feb 1, 2018 | Information, Outages, Planned Work
Further to the outage yesterday and the corrective work by the carrier we will attempt to bring the link that failed back into service this evening. If all works to plan there may be a brief network disruption while the network reconverges. If the link fails again we will remove it from the network and continue work with our supplier.
The link was brought on line approx 12:30am this morning – However the previous issue re-occurred. This has been raised to the supplier who believe this is resolved this correctly now, We will plan a further maintenance shortly to bring the link back into service. In the mean time there is no reduction to service – but there is a slight reduction in resilience to our Manchester PoP
Feb 1, 2018 | Information, Outages, Planned Work, Unplanned downtime
During planned work deemed low risk by a supplier they managed to inject a loop into one of our links. This caused a significant level of packet loss into our network at approx 22:50 on 31st January.
The link was removed from use and service was resumed albeit with reduced resilience. The issue has been reported to the carrier who have identified a potential problem and resolved this.
Jan 15, 2018 | Information, Outages, Unplanned downtime, Update
RESOLVED: 13:14pm
We are not aware of any ongoing issues now and believe that the cause of this problem has been identified and remedial action taken. Once again, apologies to anyone affected this morning.
UPDATE:
We have removed one of the backhaul lines from our network as this appears to be causing routing issues; we are seeing the majority of the affected lines coming back to their normal latency and response times.
We will continue to update here and apologise again for this affecting you at the start of the working week.
We are aware of an as yet unidentified issue affecting large numbers of our circuits leading to slow-downs, poor quality links and high latency. We are working on this now and will update here as soon as we have more information to share. We apologise that this is affecting you on a Monday morning.
Dec 4, 2017 | Information, Update
In the event that the phone system here (and the primary support number which routes via here) is unavailable for whatever reason as happened today as our VOIP system failed as part of the outage, there’s a second, fallback (geographical) number: 01480 411616, which rings directly at our support centre. this should be used.
We also suggest that all customers subscribe to our mailing list ensuring that copies of these status messages and updates are delivered to you by email or alternatively if you prefer, to your favourite RSS reader.
If you have any questions please email support@merula.net.
Dec 4, 2017 | Information, Outages, Update
UPDATE2: initial indications on further analysis are that the DDoD that was targeted at the network core which triggered one of our core routers to cease routing packets correctly. However, it did not shut-down the BGP protocols as part of this with the result that network traffic was black-holed for a large number of destinations for a period of around 30 minutes. Service was restored after the routing processes were re-started in sequence. We will continue to investigate further and update here but we believe that all services are now returning to normal. Some routers may have a stale session and will require a reboot to bring them back on-line.
UPDATE: we appear to have been the subject of a wide-spread DDoS attack (the source of which we’re currently still investigating). This caused two of our core routers to become unresponsive and this adversely affected DNS and routing outside our network. We have mitigated the attack here and all of the network is coming back on-line now. Please accept our apologies for this outage — we are aware that it affected a large section of our user-base.
We are aware of an issue affecting large sections of our network centred in London. We are working urgently to fix this and will update here as work progresses.
