Filled up filesystem

[redwings]

Hello All,

We have a filesystem that is 100% full and we are not allowed to add anymore space to it. The filesystem is just over 65 GB of data with the fragmet size at 4096 bytes. What can we do to be able to add more space to this filesystem. I know we probably could change the Fragment size, but what and how do we need to do this?

Any help is grately appreciated.

Thanks.

 

[ElgisRamon]

Have you AIX 4.3.3 with an standard journal file system? If so... you have got the absolute limit (even o AIX 5x).


I hope it works...
Unix was made by and for smart people.

 

[AIXSPadmin]

You can only select the fragment and nbpi at filesystem creation time. Either remove files or grow it are your only options.

Also, changing the fragment size would not change how big your filesystem is: if your filesystem is 4 GB with 4096 frag size you won't get more from a 4 GB with 8192 frag size. The frag and nbpi together determine how large a filesystem can be (i.e., 1TB or 50GB).

A 12 oz. box of corn flakes is 12 ounces regardless of how big the flakes are.

 

[redwings]

What do you mean I have got the absolute limit? Is there something that I can change? This is on AIX 4.3.3 with a standard file system?

 

[AIXSPadmin]

Are you out of inodes are are you out of disk space? I am assuming you are out of disk space. There is nothing that can be done except delete, compress files or add space.

The JFS restricts all file systems to 16MB i-nodes.
Maximum file system size in bytes = <npbi> * 2¬24


The JFS supports four fragment sizes of 512 , 1024 , 2048 , and 4096 byte units of contiguous disk space. The JFS maintains fragment addresses in i-nodes and indirect blocks as 28-bit numbers. Each fragment must be addressable by a number from 0 to (2 ¬28 ). The following relationship is produced:
Maximum file system size in bytes = <Fragment size> * 2¬28

Another size-related issue is the size of the JFS log. In most instances, multiple journaled file systems use a common log configured to be 4MB in size. For example, after initial installation, all file systems within the root volume group use logical volume hd8 as a common JFS log. The default logical volume partition size is 4MB, and the default log size is one partition, therefore, the root volume group normally contains a 4MB JFS log. When file systems exceed 2GB or when the total amount of file system space using a single log exceeds 2GB, the default log size may not be sufficient. In either case, the log sizes should be scaled upward as the file system size increases. The JFS log is limited to a maximum size of 256MB.

Below is the chart for AIX:

The maximum size of a JFS file system is a function of its fragment size and the nbpi value. These values yield the following size restrictions:


nbpi Fragment size in bytes Maximum size in 512-byte blocks
512 512, 1024, 2048, 4096 16777216
1024 512, 1024, 2048, 4096 33554432
2048 512, 1024, 2048, 4096 67108864
4096 512, 1024, 2048, 4096 134217728
8192 512, 1024, 2048, 4096 268435456
16384 512 268435456
16384 1024, 2048, 4096 536870912
32768 512 268435456
32768 1024 536870912
32768 2048, 4096 1073741824
65536, 131072 512 268435456
65536, 131072 1024 536870912
65536, 131072 2048 1073741824
65536, 131072 4096 2147483648

In AIX 4.3, you can have NBPI values from 512 to 128K, with corresponding
maximum file system sizes.

 

[redwings]

I guess I am pretty much out of space since I have reached 133431296 (512-byte blocks). Is there a version of AIX that does not have a filesystem limit, did AIX 5L get rid of that, or is that still going to be a problem?

 

[AIXSPadmin]

Maybe I don't understand what you are wanting? Any filesystem has a limit to the size that you make it. You cannot create a filesytem and have it dynamically grow to an infinite size.

The filesystem limits in AIX5L have been expanded with a theoretical 4 petabyte capacity and 1 terabyte tested. The inodes are dynamic in the JFS2 structure.

 

[Yegolev]

there are quite a few reasons aside from the fs limit that you should split your data amongst several smaller fs. for example, we limit our oracle datafiles to exist inside 7.5GB fs. to add another datafile requires another fs. this will limit problems in the event of lvm or disk trouble.

IBM Certified -- AIX 4.3 Obfuscation

 

[redwings]

We are currently looking at moving some of the data around into smaller fs. This is not something that was looked at when the system was first set up, now that we have reached that limit, is the time we determine how we wanted the system set up in the first place. Thanks for all the advice.