work has been a bit of a chore this week. its just been a bit busy and boring and a little annoying. i have to work with one of the guys that i dont really like. he thinks he is the shit and that he runs the place and yet it turns out he does not know nearly as much as he would like to think he does. we have another guy in our team who i by far the best techie i have ever worked with and he is exposing holes in the stuff that has been put in.
tonight we had a problem with a mail store on a mail server and it was left to me and the other guy who is really good, to sort it out. we are both contractors whilst the annoying guy is the senior member of the dept and permanent. he walked straight passed us when he saw that there was a problem and went home at 5:30 whilst we stayed behind to try and fix it, which the other guy did manage to.
i thought it summed things up quite appropriately. when the shit hit the fan he split because when things go wrong, its when you find out what people know. he did not fancy getting exposed as he has an image he is cultivating, whereas i dont give a shit if people dont think i am the best techie in the firm, and i just want to learn and try and understand things and if that means google and reading forums, then so be it. i am not too proud to admit that i dont know what every error message is. anyway, the reason the servers went to shit was because the time server on the ESX host they were on was not set right and the virtual servers were getting their time from it. 2 fuck ups and 2 more things that should not occur if you followed best practice. you dont let your virtual windows machine get time from the esx server. you use another NTP server that is not piggy backing and using an esx host as its master. in that case, even if the ntp daemon has stopped on the esx server, it still would not fuck up your guest VM's, which it did tonight. kerberos does not like time differences of more than 5 minutes between servers.
today i have mostly been listening to nina simone. class.
tonight we had a problem with a mail store on a mail server and it was left to me and the other guy who is really good, to sort it out. we are both contractors whilst the annoying guy is the senior member of the dept and permanent. he walked straight passed us when he saw that there was a problem and went home at 5:30 whilst we stayed behind to try and fix it, which the other guy did manage to.
i thought it summed things up quite appropriately. when the shit hit the fan he split because when things go wrong, its when you find out what people know. he did not fancy getting exposed as he has an image he is cultivating, whereas i dont give a shit if people dont think i am the best techie in the firm, and i just want to learn and try and understand things and if that means google and reading forums, then so be it. i am not too proud to admit that i dont know what every error message is. anyway, the reason the servers went to shit was because the time server on the ESX host they were on was not set right and the virtual servers were getting their time from it. 2 fuck ups and 2 more things that should not occur if you followed best practice. you dont let your virtual windows machine get time from the esx server. you use another NTP server that is not piggy backing and using an esx host as its master. in that case, even if the ntp daemon has stopped on the esx server, it still would not fuck up your guest VM's, which it did tonight. kerberos does not like time differences of more than 5 minutes between servers.
today i have mostly been listening to nina simone. class.
UPDATE: tonight i had to remote into one of our north american sites to shut down a server. its a small office with no IT staff and only have one server on site. now, the guy who is a bit of an asshole is a stickler for process and doing things right and getting everything documented and done by the book. well, for all that, it turns out we have a server in a remote site with no iLO remote access configured. another case of giving it large and falling over on the basics. other similar issues this week on top of the NTP issues and this were discovering they had randomly mixed 300GB and 600GB drives in the SAN for someone unknown reason, not using host sets on the SAN, not receiving alerts from the SAN, naming servers as AD domain controllers but they are not actually domain controllers, using group policy to implement Logon as a service right to a bunch of random accounts, which was overwriting the accounts that actually needed to on particular servers, incorrect IP address stated for a switch and domain controller at a remote site, so i thought they were down when i could not get to them or ping them and incorrect password written in the password list for one of the switches. so much for dotting all the I's and crossing all the t's.
No comments:
Post a Comment