i encountered that error while upgrading firmware for PERC6/I raid controller on dell’s poweredge 1950 server.
i got my hands on INTEL SSDSA2M120G2GC disk. read performance was perfect but during writes it was throwing kernel errors left and right:
once in a while i press enter too early. end result of changing permissions on the whole file system can be quite bad.
few configuration files i use on highly-available active-standby clusters running under debian.
i’ve got seemingly simple php script running every night on one of my servers… on the same machine i started to experience nightly oom-killer. i’ve found out that adodb’s Execute statement can do something completely else than you would expect from it – it loads whole result into memory rather than going row-by-row.
from this site. and it seems to be working. Go to MenuTools -> Error Console Paste the following in the “Code:” text-box: Press Enter or click Evaluate. The database will thence be vacuumed immediately.
i’m compressing quite a lot of data on the fly to save storage… until now i’ve been quite sure one cannot control compression level with oracle java’s GZIPOutputStream but apparently i was wrong.
it’s been a while since i’ve set up any [non]caching proxy servers. i’ve mostly dealt with reverse proxies, apache2 in 90% of cases, nginx – in few. below few copy&paste configs.
sometimes my backups are not consistent. rsync / samba to blame?
my old linksys router died violently – power led shines, but links on all lan ports were going down after few seconds from bootup. i bought replacement dlink dir-600 and decided to try dd-wrt firmware on it. it went better than expected!.
since a while Sphinx search can be queried not only by their own non-standard interface but also using regular mysql client. they call this feature SphinxQL.
i’ve been happy user of ESVA / global-domination [virtual in my case] appliance for mail and antivirus filtering. unfortunately it seems that was one-man project and it’s pretty much dead by now. official website – http://www.global-domination.org/ no longer works since ~ march 2010.
there is something about exchange… or at least exchange >2003. in 2008 i did my best to install exchange 2007 under windows 2008. and i failed miserably. i gave up and went for ex2003 under w2003 which serves very well. recently i tried to get ex2010 work under w2008r2. and i even succeeded after numerous ... Read More
[scratchpad] read this. debian’s bind9 works just fine. for windows 2003 server dnscmd /Config /EnableEDnsProbes 1 based on this. i got an interesting answer here
… comes but not via main entrance.
something i’ve mailed my dell sales-rep…
i’ve got new hdd delivered. it’ll be used as one of rotated disks for offline backup storage. it is WDC WD15EARS-00Z5B1 – WD Caviar Green 1.5 TB and is one of first disks with 4kB physical sectors instead of 512B ones. i’ve read at some places about performance impact of miss-aligned linux partitions for write ... Read More
goal: layer-2 failover without any special features on the switch level. after recent hang of on of switches i’d like to improve reliability of connectivity within a server rack. i already have all servers connected to two different switches… now it’s time for automated failover.
i’ve been using nagios to monitor my infrastructure for long time. in most cases it’s sending just e-mails. at ecowin/reuters i also had it connected with sms gateway. Siements TC35 connected via serial cable to linux box served us very well. i used debian’s scmxx to send messages. i’ve been thinking recently about using SIP ... Read More
i had to install system [ Debian that is ] on HP DL120G5 server.