Spectrum Storage Errors
This short guide covers most common Storage-related errors Users may encounter while utilizing the DW Spectrum. They are referred to as "Storage Failure" in the event log.
- "System disk is almost full".
Such error appears if Server detects that free space of the system partition drops under 1 GB for ARM devices and 5 GB for Desktop devices.
Spectrum considers it as an issue which should be resolved immediately due to both OS and Spectrum may suffer significantly. Although it is understandable why the OS needs free space, Spectrum needs it because it keeps important data at the system partition (e.g. system database, motion detection data indexes, audit log, etc).
- "Not enough HDD/SSD speed for recording to %drive_name%".
During the recording process, the server keeps all video frames received from every camera stream in separate buffers. The other processes of the server eventually attempt to write frames from such buffers to the physical drive.
These buffers have certain reasonable limits, independently in megabytes and number of frames.
In some cases, storage drive write-speed cannot match the rate of incoming video packets/frames which may cause a buffer overflow. When this happens the server has to drop the oldest packages. After that, the error is generated with a number of the dropped frames. User might find which camera feed specifically suffered from main media server log.
If the user receives such messages randomly from time to time with a low number of packets dropped -- there is nothing to worry about for the moment. Basically, it signals that the system is nearing its limits. So if the user is planning on expanding the system, he should expect that the number of dropped packages will increase drastically and must be resolved immediately. Options are to decrease incoming bitrate by playing with fps/image quality, replacing drives with better specs or simply adding more drives.
If the user receives such messages continuously, it should be considered critical since the footage is likely lost. In this case user should start with monitoring write disks capabilities and check their state. If everything is ok and hardware issues are not observed -- the solution should be the same as mentioned previously.
In some cases, when such errors appear, it is logical to assume that if some disk unexpectedly completely fails(local) or goes offline(network) this message will likely be the first thing user notice in the logs.
This trivially means that the Operating System returned read/write error message to Media Server on the command to read or write data block. Depending on the setup this message might pop up in pairs with the previous errors described. It definitely tells the user that major issues are happening.
E.g. Operating System met the bad blocks, the drive is unexpectedly remounted to read-only state or network file system returned remote service inaccessibility. MFT or Partition Table might be corrupted.