Martin's questions

It's well known fact that the Intel Management Engine can not be completely disabled. Its functionality can be greatly reduced using me_cleaner which utilizes various techniques including switching the ME "kill switch".

However, the Supermicro X9DRi-F BIOS contains ENABLE/DISABLE option for ME subsystem:

Motherboard manual says the following:

ME Subsystem
Select Enabled to support Intel Management Engine (ME) Subsystem, a small power
computer subsystem that performs various tasks in the background. The options
are Enabled and Disabled.
When ME Subsystem is enabled, the following items will display.
•   ME BIOS Interface
•   ME Version

When I compare the output of intelmetool -m when the ME Subsystem option in UEFI is Disabled and when the same option is Enabled, then the difference can be seen below:

$ diff -y --suppress-common-lines ~/ME_in_UEFI_disabled ~/ME_in_UEFI_enabled 
ME Status 2 : 0x39006301                                      | ME Status 2 : 0xb9000301
ME: Progress Phase          : Policy Module                   | ME: Progress Phase          : Global reset after an error
ME: Progress Phase State    : Entery into Policy Module       | ME: Progress Phase State    : Unknown 0x00
$ 

Firmware Init Complete is YES and Current Working State is Normal even when the ME Subsystem option is set to Disabled so based on this it looks that the ME is not disabled.

What does the enable/disable ME subsystem option in Supermicro X9DRi-F motherboard BIOS do?

Is it possible to configure an address from ff00::/8 address block to a NIC? Looks like it is not possible:

# ip a add ff99::1/8 dev p2p1
RTNETLINK answers: Cannot assign requested address
# 
# ip maddr add ff99::1/8 dev p2p1
"ff99" is invalid lladdr.
Error: "ff99" is not a legal ll address.
# 

Let's say, that I have a following example.com zone:

$TTL 120

@               IN      SOA     ns1     hostmaster (2018041509
                                                    300
                                                    150
                                                    600
                                                    60)

                IN      NS      ns1
                IN      NS      ns2


ns1             IN      A       192.168.0.159
ns2             IN      A       192.168.0.160

blah            IN      A       192.168.0.141

As seen above, ns1.example.com and ns2.example.com are two authoritative name-servers for that zone. When are they queried by other name-servers? As much as I tested, the only time when those two records were queried was when I asked for NS records from 192.168.0.159 with a DNS lookup utility, i.e dig @192.168.0.159 -t NS example.com.

I have an HP Smart Array P400 RAID controller with six physical disks configured in RAID 5. Four of the physical disks are in "OK" state, one is in "Predictive Failure" state, and one is in "spare" state:

logicaldrive 1 (3.6 TB, RAID 5, OK)

physicaldrive 1I:1:1 (port 1I:box 1:bay 1, SATA, 1 TB, Predictive Failure)
physicaldrive 1I:1:2 (port 1I:box 1:bay 2, SATA, 1 TB, OK)
physicaldrive 1I:1:3 (port 1I:box 1:bay 3, SATA, 1 TB, OK)
physicaldrive 1I:1:4 (port 1I:box 1:bay 4, SATA, 1 TB, OK)
physicaldrive 1I:1:5 (port 1I:box 1:bay 5, SATA, 1 TB, OK)
physicaldrive 1I:1:6 (port 1I:box 1:bay 6, SATA, 1 TB, OK, spare)

Is it safe to add (hpacucli controller slot=6 array A add drives=1:6) a "spare" drive to RAID 5 and once its drive type changes from "Spare Drive" to "Data Drive", then remove (either physically or maybe with hpacucli controller slot=6 array A remove drives=1:1) the HDD with the "Predictive Failure" state?

I have seen (older) racks which are fully-closed. For example similar with plexiglass doors:

Or with metal doors without any grid/holes. Are such racks designed for data-centers with raised floors and take cold air from the bottom and push hot air out from the top? Or are they simply for equipment which do not create (much) heat?

I have a HP ProLiant DL360 Gen9 server which should support HP Universal Media Bay. How is it connected to motherboard? If I open up the server, then there is no connector for that:

I have a SAS HDD enclosure of HP server, which has a following connector:

What is the purpose of this connector? One purpose is to lit the LEDs in front of the enclosure . Is this the only reason for that connector?

I read about registered memory and noticed that RDIMMs(and also LRDIMMs) are arced:

What is the purpose of that? Why are RDIMM/LRDIMM modules arced?

I have a HP ProLiant DL360 Gen9 server which has three physical 900GB HDDs. Those three HDDs are in some sort of hardware RAID setup. RAID controller is HP 3239 and it uses hpsa driver. For OS it represents itself as a 900GB sda device. Now if I check the RAID level with # cat /sys/class/scsi_disk/0:0:0:1/device/raid_level, then it shows RAID 1(1+0). Is this some sort of RAID 10 with three physical drives? Or is it triple mirroring? What is also interesting is that while activity LEDs of two drives blink in sync, then the third drive shows very little activity. Last but not least, I am not able to install additional software on this server because it is Linux based special appliance.

I would like to find out the authoritative name-servers for swedbank.se domain. I queried few caching name-servers for swedbank.se and while some answered the authoritative name servers, then others did not. For example Google public DNS servers:

$ dig @8.8.8.8 swedbank.se +noall +authority

; <<>> DiG 9.6.1-P1 <<>> @8.8.8.8 swedbank.se +noall +authority
; (1 server found)
;; global options: +cmd
$ dig @isp-nameserver swedbank.se +noall +authority

; <<>> DiG 9.6.1-P1 <<>> @isp-nameserver swedbank.se +noall +authority
; (1 server found)
;; global options: +cmd
swedbank.se.            2719    IN      NS      ns2.foreningssparbanken.se.
swedbank.se.            2719    IN      NS      ns1.foreningssparbanken.se.
swedbank.se.            2719    IN      NS      ns3.foreningssparbanken.se.
$ 

Is this just a matter of caching name-server configuration? Or is there some other reason for such behavior?

Lets say I create a reverse zone-file for 192.0.2.0/24 network in my DNS server and my DNS server will be the authoritative for reverse-zone 2.0.192.in-addr.arpa. In this reverse zone-file I create Start of Authority record with following parameters:

$ dig @localhost -t SOA 2.0.192.in-addr.arpa. +noall +answer
2.0.192.in-addr.arpa. 604800 IN      SOA     primary-DNS-server secondary-DNS-server 2015010600 28800 7200 1209600 86400
$ 

Am I correct that the DNS TTL for this reverse-zone is 86400 seconds? And this means that if recursive(caching) name-server asks a PTR record from this zone, then this recursive name-server will cache the result for 86400 seconds?

In case of RAID4 or RAID5, for each stripe of data bits, a parity bit is stored. For example if I write 0 to drive A and 1 to drive B, then parity bit 1 is stored to drive C. Isn't this a huge load to CPU in case of Linux software-RAID if for each bit of data, a parity bit needs to be calculated? For example if I write a 1GB file to RAID5 array, then 8000000000 XOR calculations needs to be performed by CPU?

I would like to check the availability of Internet connectivity by sending ICMP "echo request" messages to some IPv4 addresses. Those IPv4 addresses should be as reachable as possible in stability wise and they should not handle ICMP traffic as low priority. I know that DNS servers of tier-1 ISP's(e.g. Level3) and large content providers(for example Google 8.8.8.8) are a popular choice. Maybe it's wiser to use some service which is using DNS entries(IP addresses in different ISP networks) in round-robin fashion like pool.ntp.org? In a nutshell, which IP addresses do you use for this and why?

Am I correct that optimal chunk size is average file read/written to disk divided by number of block devices in RAID array storing the data? For example if the average file size is 1024KiB and I have four disks in RAID1, then I should choose the chunk size around 256KiB? Or if I have two drives in RAID0, then I should choose the chunk size 512KiB instead? Or are there better methods/benchmarks to determine the optimal chunk size for software-RAID? Last but not least, is there a good utility which could help one to measure the average I/O read/write size?

I have a /29 network of publicly-routable IPv4 addresses from my ISP. I would like to migrate to /28 network of publicly-routable addresses. Unfortunately, my ISP can not extend the current /29 network, but they gave me a new one. This means that for a next month, during the migration period, I can use both the old /29 network and new /28 network. All the web-servers have DNS A-records. What are the techniques/options to gracefully migrate web-servers from old /29 network to this new /28 network? I mean it's easy to change the network configuration in servers and restart the services, but how to ensure that web-pages are available to viewers? I guess one option would be to configure round-robin DNS in a way that domain-name resolves either to IP address in existing /29 network or to new one in /28 network and for few days, HTTP server listening on new address, just redirects to old IP address. After few days(or as long as it takes DNS records to populate across the DNS system) I would remove the old IP from DNS configuration. Will this work? Or are there better practices?

I'm aware that one can introduce packet loss in a Linux machine using tc. For example:

tc qdisc change dev eth0 root netem loss 10%

However, is it possible to specify a traffic direction where the packet loss occurs? I would like to have no packet loss on receive, but 10% packet loss on transmit for testing purposes.

"Session layer" in OSI 7-layer reference model is responsible for creating, maintaining and closing sessions between hosts. Is it possible to distinguish "session layer" in Linux network stack? What are the protocols used? For example if one downloads an ASCII text file using HTTP, then does "session layer" signal the transport layer protocols to create a connection?

What causes servers to shut down or reload in case of high temperature? For example Juniper network equipment running Junos OS will halt the routing engine if it's CPU temperature remains above 100C for more than 60s:

Jun  3 00:40:32  M10i chassisd[1209]: CHASSISD_RE_OVER_TEMP_WARNING: Routing Engine 0 temperature (101 C) over 100 degrees C, platform will shutdown in 15 seconds if condition persists
Jun  3 00:40:32  M10i chassisd[1209]: CHASSISD_RE_OVER_TEMP_WARNING: Routing Engine 0 temperature (101 C) over 100 degrees C, platform will shutdown in 15 seconds if condition persists
Jun  3 00:40:37  M10i chassisd[1209]: CHASSISD_RE_OVER_TEMP_WARNING: Routing Engine 0 temperature (102 C) over 100 degrees C, platform will shutdown in 10 seconds if condition persists
Jun  3 00:40:37  M10i chassisd[1209]: CHASSISD_RE_OVER_TEMP_WARNING: Routing Engine 0 temperature (101 C) over 100 degrees C, platform will shutdown in 10 seconds if condition persists
Jun  3 00:40:42  M10i chassisd[1209]: CHASSISD_RE_OVER_TEMP_WARNING: Routing Engine 0 temperature (102 C) over 100 degrees C, platform will shutdown in 5 seconds if condition persists
Jun  3 00:40:42  M10i chassisd[1209]: CHASSISD_RE_OVER_TEMP_WARNING: Routing Engine 0 temperature (101 C) over 100 degrees C, platform will shutdown in 5 seconds if condition persists
Jun  3 00:40:47  M10i chassisd[1209]: CHASSISD_RE_OVER_TEMP_SHUTDOWN: Routing Engine 0 temperature above 100 degrees C for too long; powering down all FRUs

However, what about servers? Does this depend on server hardware vendor? Or does this depend solely on server operating system?

At the moment I have following IPv4 addresses associated with eth0 interface:

T60:~ # ip -V
ip utility, iproute2-ss110629
T60:~ # ip addr show dev eth0
2: eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast state DOWN qlen 1000
    link/ether 00:15:58:2a:84:3e brd ff:ff:ff:ff:ff:ff
    inet 10.11.12.2/24 scope global eth0
    inet 10.11.12.3/24 scope global secondary eth0
    inet 10.11.12.4/24 scope global secondary eth0
    inet6 fe80::215:58ff:fe2a:843e/64 scope link 
       valid_lft forever preferred_lft forever
T60:~ # 

I would like to replace or change 10.11.12.3/24 address to 10.11.12.6/24. How to do this? If I execute ip addr change 10.11.12.6/24 dev eth0 or ip addr replace 10.11.12.6/24 dev eth0 then 10.11.12.6 is just added to eth0. In addition, in case of change/replace, shouldn't the syntax be change/replace from 10.11.12.3/24 to 10.11.12.6/24? I know that I can use ip addr del command, but how to achieve this with ip addr change or ip addr replace?

What is the difference between KVM and Qemu? As I understand, latter is able to provide "user mode emulation" which means that one can run a single executable built for some other architecture like this:

# file busybox-sparc 
busybox-sparc: ELF 32-bit MSB executable, SPARC, version 1 (SYSV), statically linked, stripped
# qemu-sparc busybox-sparc ls
Packages_arm     Packages_i386    arm              busybox-armv6l   busybox-i686     busybox-powerpc  busybox-sparc    busybox-x86_64   i386
# 

..and "computer emulation", which means that not only dynamic binary translation is performed, but whole virtual machine is emulated(video card, sound card, network cards, optical devices etc). For example:

# qemu-system-sparc64 -hda hda.img -cdrom debian-6.0.5-sparc-CD-1.iso -boot d

As I have understood, Qemu does not know how to take advantage of hardware virtualization(for example Intel VT-x or AMD-V solutions on x86 architecture) and this is the case where KVM can help- one can start Qemu with KVM support. However, does KVM provide any other functionality besides access for emulation programs to hardware-assisted virtualization on specific hardware architectures? Is it possible to create whole(CPU, video card, I/O devices etc) virtual machine using only KVM? If yes, then which programs should one use for creating and controlling this virtual machine?