Category computers

Note to self:

My mom’s computer was unable to check her email with Mail.app. IMAP worked fine. When logging in by telnetting to port 110, the following error about being unable to read the home directory was received, yet the proper permissions was there:

-ERR this user has no $HOME/Maildir

My mom also reported that occasionally she was able to check her email just fine.

The problem ended up being that Mail.app was set to leave all mail on the server. By the time there were around 1400 messages in her inbox (stacked up over the past eight months or so) it seems that these messages were simply too much for qmail-pop3d to parse upon login, and the error was received.

It should be noted that the log for qmail-pop3d showed a successful login with an unsuccessful one immediately following.

The solution was to ensure that she had all her mail, then clear out the POP3 server. Now everything works fine.

rez.nuxx.net (on the right) in the colo facility.

Just FYI, rez.nuxx.net is back up. Everything was done as of just before noon…

I just had the problem which normally happens when I visit the Waveform offices. That is, I get to talking with Bill there, and next thing I know a few hours have passed.

Well, I’m back home, the server is back up, and everything seems good.

To everyone who I host things for on my server, rez.nuxx.net. It seems that /dev/ad10 (one of the 120GB disks in the RAID 0+1 array which holds all the web and email data, databases, logs, aggregated backups, etc) actually has gone south. It’s still spinning and all, but when I tried to run a SMART extended offline test, there was a read error:

SMART Self-test log structure revision number 1
Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
# 1  Extended offline    Completed: read failure       20%      5330         208317256

So, I’m looking at replacing the drive sometime around 11am tomorrow, which means that the server and such will be unavailable. It’ll be back up ASAP, hopefully by noon.

Well, this is just what I wanted to see when I read my webserver’s log files this morning.

rez.nuxx.net kernel log messages:
> ad10: TIMEOUT - READ_DMA retrying (2 retries left) LBA=208317255
> ad10: TIMEOUT - READ_DMA retrying (1 retry left) LBA=208317255
> ad10: FAILURE - READ_DMA timed out
> GEOM_MIRROR: Request failed (error=5). stripe/stripe1[READ(offset=213316869632,length=4096)]
> GEOM_MIRROR: Device mirror1: provider stripe/stripe1 disconnected.

And the output from a script I wrote to check the RAID status daily:

GEOM mirror status:
Name Status Components
mirror/mirror0 COMPLETE ad0
ad2
mirror/mirror1 DEGRADED stripe/stripe0

GEOM stripe status:
Name Status Components
stripe/stripe0 UP ad4
ad6
stripe/stripe1 UP ad8
ad10

< sigh > At least I’ve got some spare disks… And I just happened to run a backup last night.

(Just FYI, there’s no data loss. One of the stripes which makes up half of the 0+1 array I use for data has gone offline due to a disk faililng.)

Gur. I’d forgotten what a pain in the ass it is to get anything working under Linux. I want to install one silly app and the requirements become a strawman diagram. :\

Problem with my laptop fixed.

First I had to edit the userinit line per KB Q249321 using the Offline NT Password & Registry Editor, then I logged in, got a blank screen, hit Ctrl-Alt-Del, fired up Task Manager, used that to launch regedt32, followed 223188 to change the boot volume back to C:, rebooted, and all was good.

The problem was caused by C: somehow changing to E: (WTF?). Yeah, I’ve got no idea how, but the problem is solved.

The error which was received is as follows:

Your system has no paging file or the paging file is too small.

This was received on my Windows XP SP2 laptop after logging into the machine with any username, then the machine would slowly return to the login screen. If one tried to log in via safe mode, they would immediately be returned to the login window.

Well, hopefully this will help someone somewhere down the line…