Reputation
7,029
Top tag
Next privilege 10,000 Rep.
Access moderator tools
Badges
15 35
Impact
~587k people reached

Jul
17
comment NFS randomly mounts wrong server
@Aaron : The FSID may be an idea. The servers are cloned, so maybe the UUID of the filesystem is identical and this causes the problems. I will see if I can set it up and test providing an FSID and see if that helps.
Jul
17
comment NFS randomly mounts wrong server
@MarkPlotnick : Nothing extra in the /etc/exports. Just the address and directory.
May
28
comment what are the minimal user permissions required to scp files to a remote server?
Yes, SCP is a standard tool. No, SCP is not a good idea if you do not want to give the user shell access on the server.
Mar
23
comment NFS randomly mounts wrong server
@MarkPlotnick: I am not on the site with the systems at the moment, so it will take a couple of weeks before I can begin debugging that again. But the pattern as far as I could tell was that the server initiated communication with the right server, and then suddenly broke off and setup the connection with another server.
Mar
14
comment NFS randomly mounts wrong server
@PaulHaldane As far as I can see there is nothing logged when mounting from one of the slaves. Where should NFS log this?
Mar
14
comment NFS randomly mounts wrong server
@PaulHaldane: The Mount output always looks correct. The incorrect servers are not reflected in the output from mount. There is no firewall in between.
Dec
16
comment cron job time managment
Note that it is usually a very bad practice to run two cron jobs where one depends on the other, as this creates a race condition. Sooner or later the second job will win the race, even if you give teh first job a 30 second head start...
Dec
15
comment Modifying kernel cmdline in Ubuntu
Gaaaa! Thanks for the hand holding! That was indeed the solution. Remind me to stab the guy who designed this mess with a rusty spoon when I get the chance.
Dec
15
comment Modifying kernel cmdline in Ubuntu
I think it's better to patch the default/grub file in this case, but editing 40_custom can certainly be an option as well.
Dec
15
comment Modifying kernel cmdline in Ubuntu
This was my first thought, but I tried exactly this on 14.01.1 (Trusty), and it has no effect. I think that the default menu options are generated from 10_linux, but regardless of if I name the file as 05_xxx or 15_xxx it is not reflected in the config generated. I know the file is sourced, as if I make a typo in it the update-grub fails.
Dec
15
comment Modifying kernel cmdline in Ubuntu
I have found the config files, but I would prefer not to modify them. Is there a way to change the default cmdline without modifying the files?
Dec
9
comment Modifying kernel cmdline in Ubuntu
@sysconfig: I kind of like that approach, but I am not sure how well uninstall/reinstall would work. Just doing a reverse patch and then barfing on uninstall if the file doesn't match seems like it could easily break. I also think that patching config files of other packages is considered bad form in Debian/Ubuntu packaging, but I may be wrong.
Apr
11
comment What's the best fire suppression for a server room?
Just a nitpick: An emergency power off does not remove fuel. EPO removes an ignition source. If you have a self sustaining fire cutting the power will not put it out (per definition), and the fuel/oxidizer/heat rule is about fighting self sustaining fires. Removing fuel is done, for example, by cutting a feeding gasoline pipe to a burning generator.
Mar
31
comment DNS resolution issue
Before you start debugging, run it through dnscheck.iis.se and make sure that you don't have some silly problem with one of the slaves...
Mar
31
comment Using pconnect on several databases in PHP on Apache2
I assume that jeffatrackaid is asking if you are using CGI or MOD_PHP...
Feb
17
comment Postfix mailqueue “Mailbox Full”
Sounds like the alarm is not tuned correctly, but it is hard to tell without knowing the use case. I would not worry so much about the length bounce queue. We do have a similar alarm, but it is configured to go off on sudden changes in the number of messages successfully delivered.
Feb
17
comment Postfix mailqueue “Mailbox Full”
@Ricardo: Is it an actual, measurable, performance problem? My experience is that there are many other things with a larger impact on the performance of the mail server than the length of the bounce queue...
Feb
13
comment Apache version 2+ doubles the TimeOut directive on requests
TimeOut is part of the low level I/O configuration in the Apache core. It's a kind of safety to prevent I/O operations that have blocked from locking the webserver. While I can't really explain the behavior it is not the normal way to control the execution time of PHP scripts. That you usually do with max_execution_time in the php.ini file.
Feb
4
comment Scheduler spikes single core on SMP Ubuntu vm under ESXi
Did you get any interesting result in your environment in the end?
Jan
28
comment Scheduler spikes single core on SMP Ubuntu vm under ESXi
Seems to have the same behavior with compression turned off, but it is not as noticeable as the dump becomes network-bandwidth bound.