Reputation
736
Top tag
Next privilege 1,000 Rep.
See votes, expandable usercard
Badges
1 4 5
Newest
 Commentator
Impact
~44k people reached

  • 0 posts edited
  • 0 helpful flags
  • 4 votes cast
Jun
22
comment How to get RX and TX bytes querying iptables?
@LivingSilver94, could you add in your exact setup and what you'd like to achieve? It might be easier to work out a solution with these facts at hand. :)
Jun
22
comment How to get RX and TX bytes querying iptables?
If your device is routing or NAT-ing, then you'll see the forward chain counters increment because packets will be traversing this chain. Remember that you can put a source and destination parameter to the forward chain rules to distinguish between LAN egress vs LAN ingress traffic.. See this article for an in-depth explanation: linux.com/learn/tutorials/…
Jun
21
comment Debian, two network cards - “no route to host” on eth1
My suspicion is that you may have an IPv4 address conflict when using the address 192.168.0.2. I'd be very keen to look at a tcpdump of traffic from the time you restart the network service to the time you get route errors. You can capture that by executing tcpdump -s0 -w route_errors.pcap -i eth1 and then pasting the output of tcpdump -r route_errors.pcap -nnn
Jun
21
comment Debian, two network cards - “no route to host” on eth1
You can instruct dhclient to inject the relevant nameserver into your resolv.conf. On my machine, I have this: ❯❯❯ grep prepend /etc/dhcp/dhclient.conf prepend domain-name-servers 8.8.8.8;
Jun
21
comment Debian, two network cards - “no route to host” on eth1
You can instruct dhclient to inject the relevant nameserver into your resolv.conf. On my machine, I have this: ❯❯❯ grep prepend /etc/dhcp/dhclient.conf prepend domain-name-servers 8.8.8.8;
Jun
21
comment Debian, two network cards - “no route to host” on eth1
Could you resolve your route -n output into addresses only (you have the name adim.hhv3)
Jun
21
answered How to get RX and TX bytes querying iptables?
May
23
awarded  Commentator
May
23
comment How to detect if USB device is bootable in udev?
You can run a script on a udev rule trigger. The trick would be for your script to scan the first few bytes of the disk for an MBR signature. The solution given by Nasha above is acceptable. Alternatively, roll your own one liner using dd and perl /grep to scan the first sector of the usb disk for an MBR signature.
May
23
comment Does anyone actually use and benefit from transparent huge pages?
The main painpoint with HugePages seesm to be NUMA related. Practically every new shiny multiprocessor server out there is now NUMA enabled. Funky things in the memory subsystem ensue once you have huge pages that cross NUMA boundaries. Have a look at: engineering.linkedin.com/performance/…
May
23
comment Docker says “no space left on device” but, system has plenty of space?
Docker is extracting a tar file from /var/lib/docker/tmp/70.... to a directory under /var/lib/docker/devicemapper/.... The fact that there's the substring 'devicemapper' in that path makes me think that docker is mapping some block storage device into a docker specific block device for use. This probably means that if you run lsblk/df and friends before or after the docker command, you'll miss the mapped device (docker cleans up after itself?). I'd propbably ctrl-z the unbundling process after the errors start popping and start poking around with df/dm tools. Can you do that?
Apr
17
comment Fixing terrible performance when watching/reading video and writing video files to disk
If you have source code access to the downloader, see whether using memory mapped files would help. I'd try: 1. Use the deadline scheduler. 2. Try to bunch up your writes into few large consecutive writes instead of many small writes (IOPS vs Throughput). 3. Tell your downloader to stop abusing the vm cache (posix fadvise). 4. ionice the downloader to a lower priority and the reader process to a higher priority. 5. Have a look at /prov/sys/vm/[dirty-ratio, dirty_bytes]. Playing with this variable will influence your writeback behaviour. linuxinsight.com/proc_sys_vm_dirty_bytes.html
Apr
17
comment Fixing terrible performance when watching/reading video and writing video files to disk
Have you tried looking at the VM subsystem hints? POSIX fadvise that is? You may find that the download process is trashing your VM cache. Download linux-ftools and wrap the download/write to disk script with fadvise dontneed and see what happens. code.google.com/p/linux-ftools
Apr
17
comment How to create a fake inotify event on debian?
Can you touch a zero sized hidden file? Say your media files are in /var/blah/media/pictures. Just touch a .touchme and that should generate the correct inotify event. Would this work?
Apr
17
comment What is the IPsec bottleneck in Linux?
Hmm, that is interesting. I'd suspect that something in the kernel is slowing things up. Would you mind sharing your ipsec configuration? I'd like to instrument a VM and see what results I'll get with different settings.
Nov
5
comment Disk is a bottleneck. Background processes make things worse. How to improve responsiveness?
YMMV: We did try to play around with swappiness unsuccessfully. We tried all kind of values, graphed stuff and didn't get anywhere. We eventually ended writing up a daemon that would evict the pages of certain misbehaving programs every few seconds from the page cache.
Nov
5
answered Dovecot/Postfix mail server replies to [SYN] with [RST, ACK] when user attempts to connect (POP3). Likely iptables?
Nov
5
answered First sudo always slow
Nov
5
awarded  Yearling
Nov
5
revised Disk is a bottleneck. Background processes make things worse. How to improve responsiveness?
added 559 characters in body