multi-master mysql replication with servers on 3 different continents

at work i’m using mysql replication quite extensively. first it was a straightforward one-way replication that has been rock-solid for us since 2009. in 2012, for another type of data, we’ve started using master-master setup. initially the servers were in different European countries, eventually the secondary site was moved to North America while primary one …

Continue reading ‘multi-master mysql replication with servers on 3 different continents’ »

E5-2643 in r620 stuck at a very low cpu frequency

few weeks ago i’ve done a general firmware upgrade of 2 identical Dell Poweredge r620 servers. a while later i’ve noticed that one of them had really sluggish performances. i’ve noticed that /proc/cpuinfo has shown CPU frequencies of hundreds of MHz rather than thousands. cat /proc/cpuinfo |grep MHz gave me 16 rows of values between …

Continue reading ‘E5-2643 in r620 stuck at a very low cpu frequency’ »

lxc broken after upgrade from linux-image-4.12.0-0.bpo.1-amd64 to linux-image-4.13.0-0.bpo.1-amd64

lxc containers no longer start after i’ve upgraded debian stretch server from 4.12 kernel from stretch-backports repository to 4.13 also from backports. symptoms in /var/log/syslog: that was resolved by creating /etc/systemd/network/99-default.link with: via this comment. then there was nothing interesting in syslog but the lxc guest still would not start – lxc-start -F -n serverName …

Continue reading ‘lxc broken after upgrade from linux-image-4.12.0-0.bpo.1-amd64 to linux-image-4.13.0-0.bpo.1-amd64’ »

btrfs and nfs dont play nicely together

i use ghettoVCB to take backups of vmware esxi 6.5 vms. data is copied to a nfs share hosted on Debian Linux. for over a year i used btrfs partition as an underlying datastore for the nfs share. and i just could not get it to work for backups which size larger than available RAM …

Continue reading ‘btrfs and nfs dont play nicely together’ »

“out of range” on drac console when booting debian stretch

i’ve installed debian stretch on an old dell poweredge 2950. after booting it i saw grub, few lines from kernel and then green screen with “out of range”: reboot, on the grub menu select “*Debian GNU/Linux”, press e, hilight a line with linux /boot/vmlinuz-4.9.0-3-amd64 root=/dev/sda1 ro quiet, add there vga=normal fb=false nomodeset and press ctrl+x …

Continue reading ‘“out of range” on drac console when booting debian stretch’ »

sphinxsearch reborn as manticoresearch?

for the past year number of commits in my favorite open source text-search server – the sphinxsearch project – has dropped to nearly 0. looks like in the meanwhile some of the old contributors moved to another project – manticoresearch. the website of the new project features two prominent supporters – craigslist and boardreader. fingers …

Continue reading ‘sphinxsearch reborn as manticoresearch?’ »

unwanted wordpress trackback spam

today i’ve learned that disabling trackback in wordpress’s settings > discussion [ ] Allow link notifications from other blogs (pingbacks and trackbacks) on new articles is not enough. your old posts, created before the date when this check-box was un-ticked, will still allow trackback which might be spammy. solution? adjust the trackback setting on per-post …

Continue reading ‘unwanted wordpress trackback spam’ »

varying svn checkout times under virtualized windows server 2016

we’ve set up a new build server using Windows Server 2016 running under ESXi 6.5. few days later devs came back with a complain: full svn checkout time varies a lot – it takes between 1 and 8 minutes. steps taken that hopefully resolved the issue: disable the icon cache of tortoise svn, then eventually …

Continue reading ‘varying svn checkout times under virtualized windows server 2016’ »

bridging two physical interfaces of esxi server

my collages got into an unpleasant situation where one of two dedicated servers, running vmware esxi 6.0, rented from a datacenter lost its network connectivity. the datacenter/internet-facing interface is down, hours later, during regular working day, the hosting provider did not react and resolve the problem. maybe the network card died, maybe switch port misbehaves …

Continue reading ‘bridging two physical interfaces of esxi server’ »

posting a file from command line with CURL

i needed to write a .bat script for windows to upload a log file. quick and nasty hack.. on the sending end: looks straightforward, worked when i tested; but failed to upload in the production setup. on the http server/PHP i was getting: turned out that the log file was growing. curl at the beginning …

Continue reading ‘posting a file from command line with CURL’ »

extracting attachments from unix mail files

i log some of the messages passing by Postfix-based relay using always_bcc = bcc@somewhere.com in main.cf. Sometimes – to track spam or problems – i need to check content of the attachments. this comes handy if content is embedded inline i have to edit the file and replace or just add the 2nd part if …

Continue reading ‘extracting attachments from unix mail files’ »

bind9 on debian jessie ignores -4 in /etc/default/bind9 and tries to make lookups via ipv6

while checking one of the servers running debian jessie i’ve noticed that logs are full of: strange… didn’t i put “-4” in the /etc/default/bind9 ages ago? apparently i did but ps shows that it was ignored: a bit of googling told me it’s a known problem with systemd [sic!] – the /etc/default/bind9 is ignored. workaround? …

Continue reading ‘bind9 on debian jessie ignores -4 in /etc/default/bind9 and tries to make lookups via ipv6’ »

sphinx search 2.2.11 – column count does not match schema (expected 3, got 2)

i have a real-time sphinx that was configured in this way:\ it was updated using REPLACE INTO myrt VALUES(?,?)… until the day i’ve upgraded to 2.2.11. the update command started to give me: after a bit of investigation i’ve discovered that: with RT index it’s obligatory to have at least one attribute [not id], even …

Continue reading ‘sphinx search 2.2.11 – column count does not match schema (expected 3, got 2)’ »

dlink DXS-1210-12TC – web interface hanging

i’ve made a mistake again and bought a ‘cheap’, smart-managed network switch. dlink DXS-1210-12TC with 12 10gbase-t ports. and i got punished nearly immediately – the web interface of the device would hang on me after few days of uptime. sometimes it wasn’t even possible to log in, other times – i could log in …

Continue reading ‘dlink DXS-1210-12TC – web interface hanging’ »

“operator or administrator has refused the request (0x800710E0)” in the windows task scheduler

at work we have a Windows 2012 server [ more than one. sadly… ]. one of them had few nightly batch jobs started from the task scheduler. a certain task would fail occasionally and give us the “operator or administrator has refused the request (0x800710E0)” error. googling did not help much – it’s a server, …

Continue reading ‘“operator or administrator has refused the request (0x800710E0)” in the windows task scheduler’ »

tuning spmassasin to treat more harshly mails with forged sender’s address claiming to come from us

once in a while we get e-mails with spoofed sender’s address claiming to come from @ourorg.com. this can fool some of our users; outlook displaying an image of sender solely based on the From: field does not help here. some of those messages have different Return-Path pointing to @someotherscammy.site, other have it also pointing to …

Continue reading ‘tuning spmassasin to treat more harshly mails with forged sender’s address claiming to come from us’ »

drac5 / 6 with latest java 1.8.73

new java – new problems. “Error when reading form SSL socket connection” showed again after the upgrade to new java follow the steps from the earlier post and edit those in the C:\Program Files (x86)\Java\jre1.8.0_73\lib\security\java.security: also: go to the Control panel > java in the General tab: Settings > Delete files – mark all the …

Continue reading ‘drac5 / 6 with latest java 1.8.73’ »

Dell’s 10gbase-t network cards on Intel’s x540 chips

we’re upgrading few servers at work to have 10gbit/s connectivity. to keep things familiar i’ve decided to stick with [power hungry and latency-inducing] cat7 copper cabling and 10gbase-t. there are few dell poweredge r620 and r630 to be upgraded. currently those have proprietary broadcom-based daughter-cards with 4x 1000base-t ports. i preferred to replace them with …

Continue reading ‘Dell’s 10gbase-t network cards on Intel’s x540 chips’ »