joe crobak's website » Apple http://www.crobak.org sharing what I find Sun, 04 Dec 2011 23:18:05 +0000 en hourly 1 http://wordpress.org/?v=3.2 Silently broken Gmail http://www.crobak.org/2011/04/silently-broken-gmail/ http://www.crobak.org/2011/04/silently-broken-gmail/#comments Sat, 02 Apr 2011 22:52:04 +0000 joecrow http://www.crobak.org/?p=60 Continue reading →]]> At work, we have google apps, which comes with several gigs of gmail storage. For email, though, we use outlook server with a low quota. Rather than deleting email, I “archive” to gmail via IMAP.

One day, though, gmail IMAP silently stopped syncing. I could login, but no mail was being transfered. I tried everything — synchronizing accounts, rebuilding the mailbox, but nothing worked. I was hoping to enable some more verbose logging, when I came across an old article from eriklabs.com about Mail.app Logging.

After learning the magic incantation, I started up Mail.app to log all operations on port 993 (the SSL port that gmail IMAP uses).  A message in the logs immediately stood out:

5.1 BAD [ALERT] Message too large. http://mail.google.com/support/bin/answer.py?answer=8770

The prior message logged the timestamp of the offending email, which was over 25MB in size (the max size for Gmail).  Unfortunately, Mail.app had created a large number of copies of this email (over 1,000) and placed them in the “Recovered Items” folder.

Even after deleting all of these “recovered item” copies, Mail kept making new one.  I tracked down a copy of this file in the .OfflineBackups directory. After removing the file that gmail rejects (see http://automatica.com.au/2010/01/mail-app-and-its-offline-cache/ — I chose the file with the large file size in that directory), I was finally able to resync with gmail!

Unfortunately, mail.app was not happy that I removed a file from .OfflineBackups, and it refused to process the rest of the files in that directory.  I wrote the following python script to convert the unprocessed files in .OfflineBackups into a single .mbox file suitable for import into Mail.app.

Thankfully, in the end, everything is back to normal.

]]> http://www.crobak.org/2011/04/silently-broken-gmail/feed/ 1 JAVA_HOME on Mac OS X http://www.crobak.org/2009/10/java_home-on-mac-os-x/ http://www.crobak.org/2009/10/java_home-on-mac-os-x/#comments Fri, 09 Oct 2009 00:57:50 +0000 joecrow http://crobak.org/?p=28 Continue reading →]]> I was working on configuring HBase to run on my Mac OS X machine, and I ran into a hiccup setting up the JAVA_HOME environment variable. Eventually, I determined that there’s a “Home” directory inside of each Java Framework. So, the full command is is:

export JAVA_HOME=/System/Library/Frameworks/JavaVM.framework/Versions/1.6.0/Home

]]> http://www.crobak.org/2009/10/java_home-on-mac-os-x/feed/ 0 Globetrotter Connect for Option PCI Express card and Mac OS X 10.5.7 http://www.crobak.org/2009/06/globetrotter-connect-for-option-pci-express-card-and-mac-os-x-1057/ http://www.crobak.org/2009/06/globetrotter-connect-for-option-pci-express-card-and-mac-os-x-1057/#comments Mon, 08 Jun 2009 22:53:00 +0000 joecrow http://crobak.org/?p=8 Continue reading →]]> Edit: I may have jumped the gun.  After a few more days of use, it seems like the Globetrotter Connect software is no more buggy than it was before the 10.5.7 update.  The multiple kernel panics must have been a coincidence.

Warning! I installed the Mac OS X 10.5.7 update this past weekend, and I’ve had a lot of problems today with the Globetrotter Connect software. It has caused 2 kernel panics in under an hour of use :-( . If you use this software, I’d strongly suggest waiting for a new version before upgrading to Mac OS X 10.5.7.

In a few days, I’ll give an update on if the KPs continue, or if it’s a transient error.

Here’s the lovely KP back trace that I’ve seen a numer of times:

Mon Jun  8 08:48:18 2009
panic(cpu 1 caller 0x003A04EC): "m_free: freeing an already freed mbuf"@/SourceCache/xnu/xnu-1228.12.14/bsd/kern/uipc_mbuf.c:3574
Backtrace (CPU 1), Frame : Return Address (4 potential args on stack)
0x3b13b7f8 : 0x12b4c6 (0x45ec20 0x3b13b82c 0x13355c 0x0)
0x3b13b848 : 0x3a04ec (0x498e5c 0x467ae00 0x3b13b878 0x3f0864)
0x3b13b8c8 : 0x34a2bff2 (0x3072c200 0x3072c200 0x0 0x5a95fb)
0x3b13b8f8 : 0x5a21cb (0x5cb3c00 0x0 0xe00002eb 0x3e)
0x3b13b968 : 0x811489 (0x43c6000 0x4679500 0xe0004050 0x3e)
0x3b13b9d8 : 0x80f17e (0x43c6000 0x4671d00 0xe0004050 0x4673c00)
0x3b13ba38 : 0x80f1f1 (0x43c6000 0x2 0x3 0x0)
0x3b13ba68 : 0x4268e5 (0x43c6000 0x2 0x3 0x0)
0x3b13bab8 : 0x5a2472 (0x464d4c0 0x59e99a 0x2 0x3)
0x3b13bae8 : 0x5a6d08 (0x43c6000 0x2 0x4991ccc 0x4996d40)
0x3b13bb08 : 0x34a2bb6b (0x4991cc0 0x49eef0c 0x3b13bb38 0x5cb3c00)
0x3b13bb28 : 0x34a2bdec (0x5cb3c00 0x0 0x0 0x0)
0x3b13bb58 : 0x5e605e (0x5cb3c00 0x5e84800 0x5809e20 0x0)
0x3b13bb88 : 0x5e5e88 (0x5cb3c00 0x2 0x5e84800 0x0)
0x3b13bbc8 : 0x4268e5 (0x5cb3c00 0x3b13bc40 0x0 0x0)
0x3b13bc18 : 0x5e5f04 (0x498dd80 0x5e5e46 0x3b13bc40 0x0)
	Backtrace continues...
      Kernel loadable modules in backtrace (with dependencies):
         com.option.driver.OptionQC(1.1.18)@0x34a26000->0x34a30fff
            dependency: com.apple.iokit.IOSerialFamily(9.4)@0x69e000
            dependency: com.apple.iokit.IONetworkingFamily(1.6.1)@0x5de000
            dependency: com.apple.iokit.IOUSBFamily(3.4.3)@0x59d000
         com.apple.driver.AppleUSBUHCI(3.3.5)@0x80b000->0x818fff
            dependency: com.apple.iokit.IOPCIFamily(2.6)@0x5cd000
            dependency: com.apple.iokit.IOUSBFamily(3.4.3)@0x59d000
         com.apple.iokit.IOUSBFamily(3.4.3)@0x59d000->0x5c5fff
         com.apple.iokit.IONetworkingFamily(1.6.1)@0x5de000->0x5f3fff

BSD process name corresponding to current thread: kernel_task

Mac OS version:
9J61
]]> http://www.crobak.org/2009/06/globetrotter-connect-for-option-pci-express-card-and-mac-os-x-1057/feed/ 0