A bright and diverse world of HDD failures

Yesterday there was a support request on another weird problem which was eventually forwarded to me for investigation. The symptom was “application configured as Windows service does not start and shows “Starting” in services console all the time”. The case did not look difficult as the symptom clearly indicates a freeze or deadlock inside service, but the task is anyway to find the reason and make it fast.

The symptom was exactly the same if software was started as a regular application. Luckily I already have the tool available, ProcessSnapshot, to troubleshoot things easily. The call stack indicated that software froze inside CreateFile function and further investigation revealed that it was a particular drive which was causing the problem. This was a digital video recording server with several HDDs attached building up a total video archive capacity of over 7 terabytes.

So basically the job was done but what was particularly interesting: the drive was still accessible from Windows Vista shell explorer, however there was a noticable delay opening a subdirectory. I was trying to find an evidence of HDD failure in system event logs but there were none. The drive did looks healthy in system information and logs, shown unreasonable delay but still was openable when browsing directory, and was completely freezing an application which attempted to write to this drive. And as it is natural in case of such symptoms, the first suspect is software and it was us to find the real reason.

… I appreciate all the support. It’s working very nicely again. I think I have a faulty 1.5 Tb drive (caution with those drives – I heard on the news that they’re failing).

By the way, the drive was Seagate Barracuda 7200.11 SATA 3Gb/s 1.5-TB Hard Drive (ST315003)

Leave a Reply