It's a perfectly stupid idea, I know. But I want to test one of my older projects, which includes kernel module (Netfilter extension) on a virtualized cloud-y instance.
Will it be possible for me to plug kernel code into Amazon's virtualized (I see they're using Xen.. instances?
(I think, if preparing an AMI for HVM includes a kernel configuration to support Xen drivers, then it must be possible to load a "normal" module..)
What is the crontab syntax for specifying running hours/minutes based on the day of the week in a single entry?
There's a task which needs to be executed every /1 minute from Monday,01:00 to Saturday,01:00, so after specifying the days (1-6) I still have to exclude 00:00-00:59 from Monday and 01:00-23:59 from Sunday.
I don't want to keep it done with three entries instead of one; doing additional time check via some wrapper script is also not a good option.
Cron flavor: vixie-cron, could be changed to another if required.
I'm digging through various Redis/Ceph/Elliptics benchmark to determine, whenever its theoretically possible to reach (very optimistic) request completion time below milliseconds in a non-SSD non-million worth-of-SAS-disks environment.
It is known that for localhost installations, random read performance for Elliptics large dataset is around 3.5ms/item, and almost same time is officially given for Redis.
(Note: network latency is ignored for this question).
Is there a known K-V store with higher confirmed performance?
I wanted to download some backups (about 4..5Gb in size) via HTTP and was surprised that IIS7 didn't allowed me to do so. The attempt was seemingly blocked - got a 404 page.
Where should I go in the IIS Management console to allow such downloads?
I'm setting up a link tracking system, which (temporarily) involves adding /link/id/ in front of URL (like http://server/data/id/publication/id/).
rewrite data/id/(.*) http://server/$1;
The request is logged as:
ip - - [17/Nov/2011:10:07:19 +0300] "GET /data/id/publication/id.html HTTP/1.1" 302 154 "-" "UA"`
For some reason (keeping the compatibility with AWStats) it is wanted to have 200 logged instead of 302. (nginx allows to get 301 code out of box with permanent option, but thats inappropriate too)
What are my options here?
Will the combination of location { } and rewrite do the job?
UPDATE: AWstats is able to understand various status codes by setting ValidHTTPCodes="200 304 301", see this question; my problem is solved, but question is still active - any way to get rewrites working without 301/302 code?
I need to make a copy of a working linux system, which (being contained in /, and the new place is prepared in /mnt/sdb5) seems to contain considerable amount of hard-, soft-links and special files in /dev; would cpio handle this job without applying additional magic?
Currently known safeguards:
root=/dev/sdb5; before removing, take a full partition backup.But still, not going to lose time just because cpio missed some flag and crippled the permissions/node type/soft or hardlink.
Which tool to use / which underwater rocks to avoid?
I need to change some hosts entries - for example, there's a desire to update hosts to have active.vm host match the IP of currently booted VM (with some testing appliance). Guest software within VM reports machine's address back to host, and I need to update /etc/hosts based on that information - there's always some time for writing another daemon/cronjob, but wanted to use something existing for fixing hosts entries with a dumb syntax of hostname current_ip.
Someone had recommended me the bind-tools package, but - its nsupdate|host don't suit the task at all. ..
Upd: still looking for something from bind suite, but at least dynamic regeneration seems to be the best solution, and augtool is nice too.
Upd: due to my laziness, dynamic regeneration is not planned to use right now, so started using augtool at the cronjob, that reads webserver's access logs for latest IP - at the VMs which run Windows, whole monitoring solution is just a webbrowser which loads specific url from host machine and has a small reload interval
my unfortunate Opera lost the notes again -
There's an alternative to iptables (command) which was able to automatically clean unused rules, support rule blocks etc. No, its not EBTables.
I need to add rules which will get disabled by it's rule manager / parser after given timeout, and know that such software exists; so no wheel reinvention expected - just help me remember the name..
I need to measure IPTables/inline Snort impact on performance, so need to put my test server under a good load.
Unfortunately, no hardware traffic generator is available, thus connections must be real - e.g. come from other machines (have some in same network) and use some daemon. It's not like sending 1000's of packets to see how they go rejected and dropped, the connections must be over a minute long and carry some data.
Any good solutions to simulate a thousand of long connections from a Linux machine?
Task: run blocks consisting of 3-5 commands (in parallel/background). Example block:
dd if=/dev/urandom of=/mnt/1/x bs=1024 count=1024000000 &
dd if=/dev/urandom of=/mnt/2/x bs=1024 count=1024000000 &
dd if=/dev/urandom of=/mnt/3/x bs=1024 count=1024000000 &
When it's done, next block should run. I suppose, this can be done via lock files:
task1.sh:
real_task1 real_param1 ; rm /var/lock/myscript/task1.lock
task2.sh:
real_task2 real_param1 ; rm /var/lock/myscript/task2.lock
...
taskgen.sh:
# loop
# while directory isn't empty - wait...
gen_tasks.pl # build task files from some queue
for i in 1 2 3; do touch /var/lock/myscript/task$i.lock ; done
./task1.sh &
./task2.sh &
./task3.sh &
# if task1.sh doesn't exits then exit, else loop waits for files to be deleted
A number of methods to check if the directory is empty can be found here, don't sure which to use;
Question: any better way to implement this?
P.S. Possible status reporting method:
command && report_good_state.sh taskid ; report_state_done.sh taskid; rm /var/lock/myscript/taskN.lock
It is known that certain AMD64 Linuxes are subject of being unresponsive under heavy disk I/O (see Gentoo forums: AMD64 system slow/unresponsive during disk access (Part 2)), unfortunately have such one.
I want to put /var/tmp/portage and /usr/portage trees to a separate partition, but what FS to choose for it?
Requirements:
* for journaling, performance is preffered over safe data read/write operations
* optimized to read/write 10000 of small files
Candidates:
* ext2 without any journaling
* BtrFS
In Phoronix tests, BtrFS had demonstrated a good random access performance (fat better than XFS thereby it may be less CPU-aggressive). However, unpacking operation seems to be faster with XFS there, but it was tested that unpacking kernel tree to XFS makes my system to react slower for 51% disregard of any renice'd processes and/or schedulers.
Why no ReiserFS? Google'd this (q: reiserfs ext2 cpu):
1 Apr 2006 ... Surprisingly, the ReiserFS and the XFS used significantly more CPU to remove file tree (86% and 65%) when other FS used about 15% (Ext3 and ...
Is it same now?
Wanted to move a bunch of VPN configs (including vpn-via-vpn-via-vpn scenarios) to existing Win7 installation under VMware. So the actual title is: "IPRouterEnable in virtualized Windows 7: how to redirect traffic, coming to VM, to VPN connection open at VM?"
There's a NAT configured at vmware, network 172.16.12.0, the experimental Win7 VM at 172.16.12.128 can connect to VPN, is available from host os and can run a proxy service available from outside.
Next, wanted it to route request;
How can I tell Win7 to forward all incoming requests via established VPN connection? Or, is this a routing problem in host os (Linux)?
I have a motherboard with onboard NIC (single Ethernet NIC in the computer); also USB port is constantly occupied with Wifi dongle.
in my /etc/udev/rules.d/70-persistent-net.rules, found this:
# PCI device 0x10ec:0x8168 (r8169) SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="6c:f0:49:e8:22:cd", ATTR{dev_id}=="0x0", ATTR{type}=="1", KERNEL=="eth*", NAME="eth0"
# PCI device 0x10ec:0x8168 (r8169) SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="00:1e:68:dc:96:52", ATTR{dev_id}=="0x0", ATTR{type}=="1", KERNEL=="eth*", NAME="eth1"
Currently, eth0 is detected. If I reboot, it may become eth1 (don't sure if mac will change, I wont reboot this PC today).
# ifconfig
eth0 Link encap:Ethernet HWaddr 00:1e:68:dc:96:52
Sometimes, eth0 is not detected and eth1 comes up. Why? Can it be some BIOS bug with setting onboard MAC when cold-booted improperly?
Currently the ruleset is changed to force name to eth0 (for both macs), should the next reboots reveal if it helps..
Will this rules make kernel to see the same packet at least twice?
iptables -t nat -A POSTROUTING -o eth0 -j MASQUERADE
iptables -t nat -A POSTROUTING -o ppp0 -j MASQUERADE
Found the stuff in a forgotten config file. A remark about option "-d 0/0" was also there, did this option makes sense when certain log targets are used in iptables - to avoid incorrect traffic calculation?
I currently have this snippet:
# flush all chains
iptables -F
iptables -t nat -F
iptables -t mangle -F
# delete all chains
iptables -X
Is there a possibility that some impervious rule will stay alive after running this?
The idea is to have a completely clean iptables config, that can be easily replaced by new ruleset (nevermind routes/ifconfig's parameters).
how much TCP packets should come at 100/1000mbit network card (not an embedded Realtek's home solution) before the Linux kernel will become unable to process the NIC's buffer resulting in packet drop?
Hardware: let's say about dual-core Xeon + Intel's gigabit PCIE card. The final product is planned to capture and process the packets coming from a large net (doing the same job that Netflow does, but with few specific features). But currently, have no test environment available to pre-calculate possible load per server.
Software for such a basic load test: linux's kernel ~2.6.30, netfilter, 2-3 iptables rules. Some kind of network similation is in progress, too.