Kilo

1,099 reputation

512

bio	website
	location
	age
visits	member for	3 years, 11 months
	seen	May 14 at 2:55
stats	profile views	100

	1,099 reputation	bio	website		visits	member for	3 years, 11 months
512 badges		location			seen	May 14 at 2:55

summary answers questions tags badges favorites bounties reputation activity

56 Comments

suggestions reviews revisions comments badges posts accepts all

Feb 14	comment	GLUSTERFS or DFSR Looked but can't find the article. Been running DFSR since I made this post & it has been very hands-off. IE it just works. One thing that got me into a little trouble was forgetting to set priority on DFS shares. I keep data on both servers and run full mesh. Some apps access data on srvr1 while other apps use srvr2 and so priorities should be set to prevent workstations from using less preferred node for a given app. Being a huge Linux advocate, I must say it was nice not mucking around in .conf files and ./configure;make;make install; not typical of M$ to work this well out of the box.
Aug 12	comment	OpenVPN not Routing correctly on CentOS 6 you have two default routes with the preferred (lower metric) pointing to something other than your VPN interface. I would fix this problem if you want all traffic to be routed over the VPN when it's up.
Aug 11	comment	OpenVPN not Routing correctly on CentOS 6 is windows firewall on? Why are you using a well known port (443) not that this is a problem but seems odd. What does "route print" and "ipconfig" show on your windows client? Does your default GW IP match your TUN address?
Aug 10	comment	Win2K8 R2 Scheduled Task Fault Tolerance Hopeless: My thoughts exactly. I reckoned there has to be a cleaner solution though? One of the problems I forsee is the attempting to fence the correct server to allow the scheduled task to complete on which sever (assuming if I launch it on both servers at same time). My other thought was to just stagger the jobs in time so that the DFS replication is not trying to resolve concurrent changes to the replicated space.
Aug 10	comment	kill -9 a postgres process @Craig: What a superb response; and to include a demonstration, I wish I could up vote this 100x. I'm a software developer who works with PG daily and since the 6.x days and your response is spot on! Nice!
May 25	comment	GLUSTERFS or DFSR Ah, very cool. Like you I tend to goto Linux as my first choice but two things causing me to look closer to Windows besides DFSR. (1) Customer wants centralized Antivirus and the Windows product just seems painless. (2) Some workstations have roaming profiles and they don't always roam; always assumed incompatibility with Samba being domain controller. My thought is to put Samba and GLUSTERFS on same host and if the main server fails, I can fail over Samba and would hope GLUSTERFS would just take care of things.
Oct 3	comment	100% uptime for a web application I beg to differ. "Not useful at all?" It was most certainly useful for me and contrary to your "hire someone else to do it" comment, I suppose with your reasoning the guy should trench his own fiber optic cable and design his own switches rather than buy them too? Are you serious, Yves? You sound like someone who has not spent much time in the IT field.
Apr 8	comment	Using NTP to sync a group of linux servers to a common time source @MattBianco: No. We use a custom script for that. It is an implementation detail specific to the environment. To shed some light on this, the script runs on our masters and requests the time from each client node out there. It knows of the clients as all our hosts use central LDAP. This script is part of many we run as a daily "morning report" we call it. Variables in the script allow us to set thresh holds which we report out through a nagios dashboard.
Mar 11	comment	Is it normal to get hundreds of break-in attempts per day? while I agree with others this is normal on common ports required (80, 443) I practically eliminated these attempts against my SSH port by simply changing the default port from 22 to something obscure like 6022 for example. Just doing that, alone, nearly eliminated 99% of that type of attack.
Feb 18	comment	Best VoIP VPS Location? QoS does you no good against in-bound traffic, keep in mind. It's good to have but it's not a complete solution by itself. The answer, however, to #1 should be obvious. Lower latency is always better.
Dec 25	comment	Apache RewriteRule with a RewriteMap variable substitution for the VAL argument to environment variable I solved it. I overlooked the fact that the md5digesthex.pl had to be a long running app. So I change the perl code run inside a while(<STDIN>) loop instead and I now get the desired results... Thanks for the help.
Dec 24	comment	Apache RewriteRule with a RewriteMap variable substitution for the VAL argument to environment variable Yes, wish I was around to slap the original developers. Anyway, the output of the md5inhex is just a hex string. I pass it the name of a file, it computes the MD5 and returns the hex value i.e. 0xBADBEEF.... I've tested it on the command line and it works and a browser test prompts me to download the file but HTTP header never gets set.
Dec 24	comment	Apache RewriteRule with a RewriteMap variable substitution for the VAL argument to environment variable very true. I guess I need to explain why I need to solve this? OK, it's a very large system. Some 4 or 5 back end systems that all use the same MD5 in hex format and the client is at some 30,000 locations. When you have a very large system such as this, you simply can not change the client and be done. This is a tactical approach until a more strategic architecture can be developed. We currently serve up binaries from a database application and the load is growing too high hence the reason to move to a web server.
Nov 20	comment	VPN connection over SSH well, plugin the IP given to you by your DHCP server. Just know that if your IP changes, you'll need to adjust accordingly. If you leave your machine on then your IP should renew and stick unless your DHCP provider purposely forces new IPs upon you.
Nov 20	comment	Using NTP to sync a group of linux servers to a common time source We have over 700 servers and VMs in our 3 data centers and none are off by more than 1 or 2 seconds. Anything beyond 1-2 seconds is almost always due to a recent reboot, etc... Normal daily operations for us is everything sub-second synchronized.
Nov 20	comment	Using NTP to sync a group of linux servers to a common time source Yes, we monitor using a number of custom scripts. Some scripts run on the clients that just make sure the NTPD daemon, among other things, is running and if it is found not running will attempt to restart it and if that fails, thing are logged and emails are sent. Data center time is tracked by external systems - a combination of Nagios and Munin for the most part.
Nov 19	comment	Apache 2.2 Open SSL wildcard Do you know if your certs need intermediate certs? The example given shows no intermediate certs and many 'new comers' to the SSL cert offering market tend to require this. Trustico mentions GeoTrust and Verisign so it would appear they are an intermediary. Post a https URL that is protected by one of their certs and I'll examine it for you (if you have one).
Nov 19	comment	Apache 2.2 Open SSL wildcard "each different IP" If you have a separate IP for each FQDN, then why are you using wild card certs? (cost savings or something?) Yes, I believe you need the UseCanonicalName On directive.
Oct 27	comment	Server crashing a lot of times every day. How to find what is the probem Is this a physical server or is this Linux within a VM? If physical have your ruled out faulty memory chips? If this is a VM, then do the other VMs go down as well.
Oct 26	comment	Microsoft Exchange rejecting logins Thanks, that helped fix it. The ExchangeRPC and ServiceHost were not running as were a couple others. I guess I over looked this when I rebooted the box and didn't see any Event logs pertaining to failed services. The Test-SErviceHealth help me identify required services that were not running.