239 reputation
216
bio website identi.ca/grahamperrin
location United Kingdom
age 49
visits member for 3 years, 4 months
seen Sep 16 at 18:36

Day job: IT support to CENTRIM and Brighton Business School – four days per week at Freeman Centre on University of Sussex campus, one day at Mithras House on University of Brighton campus.


Oct
8
revised For L2ARC and ZIL: is it better to have one large SSD for both, or two smaller SSDs?
A little more specific … a question that's ultimately "about a white line in a ZFS diagram" may be not a good fit with Server Fault, but already there's a good answer …
Oct
8
revised For L2ARC and ZIL: is it better to have one large SSD for both, or two smaller SSDs?
Background.
Oct
7
awarded  Student
Oct
7
comment For L2ARC and ZIL: is it better to have one large SSD for both, or two smaller SSDs?
I see two current votes to close, instead can we simply move the question? Thanks @afrazier
Oct
7
comment Disk IO slow on ESXi, even slower on a VM (freeNAS + iSCSI)
Part of this answer might be good under a question that's currently in Super User: For L2ARC and ZIL: is it better to have one large SSD for both, or two smaller SSDs?.
Sep
24
asked For L2ARC and ZIL: is it better to have one large SSD for both, or two smaller SSDs?
May
21
awarded  Autobiographer
Aug
18
comment Defragmenters useful for *nix?
If TN1150 (March 2004) is up-to-date, JHFS+ now is no different to JHFS+ then … however the more modern Systems seem to use the format in different ways. I plan to post an answer under Do Macs need to be defragmented? covering at least some of this. Meantime … rather than discuss in comments, let's aim for Ask Different Chat with the option of bookmarking.
Aug
18
comment Defragmenters useful for *nix?
It was an excellent article for its time, but the seven years since then have seen significant changes to preferred disk capacities, to uses of JHFS+, to uses of the attributes B-tree et cetera. The article does not consider fragmentation of attributes B-tree and (more commonly) catalog B-tree files; concerning the latter there's an answer to an Ask Different question relating to slowness of Time Machine. Considering modern use cases, I believe the 2004 conclusion is outdated.
Aug
18
awarded  Excavator
Aug
18
comment Defragmenters useful for *nix?
The author begins with a note: "… sampled too few a volume to generalize my "results" …".
Aug
18
awarded  Editor
Aug
18
comment Defragmenters useful for *nix?
The 25 June 2009 edition of support.apple.com/kb/TS1417 is misleading. It mentions fragmentation in the context of fsck but fsck_hfs does not give information on fragmentation. Older versions of fsck_hfs have a strict requirement for contiguous free space for one type of rebuild; in other words: if free space is too fragmented, then rebuild is impossible. More recent versions of fsck_hfs are less demanding but still, it's sane to have plenty of contiguous free space.
Aug
18
revised Do I need to defrag MacOSX filesystems?
The System can defragment only *some* things. Benefits of restarting may be debatable; I have added my interpretation of the likeliest benefit.
Aug
18
suggested approved edit on Do I need to defrag MacOSX filesystems?
Aug
18
comment Do I need to defrag MacOSX filesystems?
The gist of the answer "… you may benefit …" gets my vote up. The article referenced has been archived and is no longer updated by Apple. (February 2010 edition removed a reference to another archived article; it's otherwise no different from the June 2008 edition (crawled around that time) — not applicable to Snow Leopard Mac OS X 10.6 or Lion Mac OS X 10.7.
Aug
18
comment Do I need to defrag MacOSX filesystems?
The conclusion is reasonable based on results, but the author begins with a note: "… sampled too few a volume to generalize my "results" …".
Aug
18
awarded  Supporter