Re: Holding disk size misread by amcheck
LaValley, Brian E wrote: Good point, my tape drive maximum sustained data transfer rate is 60 MBytes per second on a Gigabit ethernet network. Is that too slow? Maybe. Maybe not. Those modern drives sometimes (usually) slow down their motors when the bytes do not flow in fast enough, so that they keep streaming, but only at a lower speed. Even gigabit ethernet cannot do 120 Mbytes/sec, or 60MB/sec full duplex. To find out how amanda gets the numbers, you may download the sources of amanda, do a complete configure and compile ("make install" not needed), and create a little standalone program "statfs", like this: $ ./configure --with-user=... etc. (not necessary to get everyting completely correct, you don't need to do a "make install"; compile should succeed, of course) $ make $ cd common-crc $ make statfs And now run the standalone "statfs" on those filesystems. It will show what kind of statfs call it uses, and the values of the different fields the OS gives back. e.g. $ /tmp/quick/amanda-2.4.5/common-src/statfs . statfs (SVR4 (Irix-5+, Solaris-2, Linux glibc 2.1)) name totalfree avail files ffree favail --- --- --- --- -- -- -- . 129015488 83820512 77266912 16384000 16137345 16137345 As you can see, there is sometimes a difference between "free" and "avail". I'm not sure why... probably the reserved space in the filesystem. Amcheck uses the "avail" value. My "df" is consistent: $ df -k . Filesystem 1k-blocks Used Available Use% Mounted on xlate:/space/tm 129015488 45194976 77266912 37% /net/xlate/tm Amcheck produces: Amanda Tape Server Host Check - WARNING: holding disk /backup/amanda/dumps/dump2: only 88344348 KB free (104857600 KB requested) WARNING: holding disk /backup/amanda/: only 8842796 KB free (26214400 KB requested) amcheck-server: slot 0: date Xlabel DailySet1-A00 (new tape) NOTE: skipping tape-writable test Tape DailySet1-A00 label ok Server check took 690.657 seconds Amanda Backup Client Hosts Check Client check: 1 host checked in 0.063 seconds, 0 problems found (brought to you by Amanda 2.4.4p2) The NFS mount is /backup/amanda/dumps/dump2. "df" produces: Filesystem 1K-blocks Used Available Use% Mounted on jhaas-l:/native/amanda 207955672 25107104 172285032 13% /backup/amanda/dumps/dump2 So in reality, there is 172GB free and Amanda only sees 88GB free. What is the 32 bit long size you mention in amanda? Amanda internally uses a signed 32bit long, which holds the size in units of 1 Kbyte. That means such variable can hold values up to 1024 * 2^31 bytes, or about 2 TByte. -- Paul Bijnens, XplanationTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, ^^, * * F6, quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... * * ... "Are you sure?" ... YES ... Phew ... I'm out * ***
RE: Holding disk size misread by amcheck
Good point, my tape drive maximum sustained data transfer rate is 60 MBytes per second on a Gigabit ethernet network. Is that too slow? Amcheck produces: Amanda Tape Server Host Check - WARNING: holding disk /backup/amanda/dumps/dump2: only 88344348 KB free (104857600 KB requested) WARNING: holding disk /backup/amanda/: only 8842796 KB free (26214400 KB requested) amcheck-server: slot 0: date Xlabel DailySet1-A00 (new tape) NOTE: skipping tape-writable test Tape DailySet1-A00 label ok Server check took 690.657 seconds Amanda Backup Client Hosts Check Client check: 1 host checked in 0.063 seconds, 0 problems found (brought to you by Amanda 2.4.4p2) The NFS mount is /backup/amanda/dumps/dump2. "df" produces: Filesystem 1K-blocks Used Available Use% Mounted on jhaas-l:/native/amanda 207955672 25107104 172285032 13% /backup/amanda/dumps/dump2 So in reality, there is 172GB free and Amanda only sees 88GB free. What is the 32 bit long size you mention in amanda? -Original Message- From: Paul Bijnens [mailto:[EMAIL PROTECTED] Sent: Wednesday, August 24, 2005 4:20 AM To: LaValley, Brian E Cc: Amanda (E-mail) Subject: Re: Holding disk size misread by amcheck Paul Bijnens wrote: > LaValley, Brian E wrote: > >> I have a Fedora Core 3 amanda server. I have specified an nfs mounted >> directory for one of the holding disks. Does anyone know why amcheck >> finds >> much less space available on this drive than a command like 'df' does? >> > > You need to explain a little more. "amcheck" checks only for > free space (at least 64K) n AMANDA_TMPDIR, AMANDA_DBGDIR, and /etc. > "amcheck" does not check space on the disklist entries. > Show the messages that make you that. > Oops, following up on myself. Still no coffee here, only water. Holdingdisk is indeed checked by amcheck, sorry. Two remarks: NFS is usualy a bad choice for holdingdisk. Can your network feed the bytes to the tape at tapedrive speed? Calculate the bandwith needed for the incoming bytes too. If you cannot feed the bytes fast enough to a tapedrive, the drive needs to stop/rewind/restart frequently, loosing lots of capacity, time, and will need replacement very soon. About the stated capacity: could it be you're somewhere in the terabytes of free space, and you're overflowing a 32bit long (size in KB used inside amanda)? Showing the exact output of the commands still helps a lot. -- Paul Bijnens, XplanationTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, ^^, * * F6, quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... * * ... "Are you sure?" ... YES ... Phew ... I'm out * ***
Re: Holding disk size misread by amcheck
Paul Bijnens wrote: LaValley, Brian E wrote: I have a Fedora Core 3 amanda server. I have specified an nfs mounted directory for one of the holding disks. Does anyone know why amcheck finds much less space available on this drive than a command like 'df' does? You need to explain a little more. "amcheck" checks only for free space (at least 64K) n AMANDA_TMPDIR, AMANDA_DBGDIR, and /etc. "amcheck" does not check space on the disklist entries. Show the messages that make you that. Oops, following up on myself. Still no coffee here, only water. Holdingdisk is indeed checked by amcheck, sorry. Two remarks: NFS is usualy a bad choice for holdingdisk. Can your network feed the bytes to the tape at tapedrive speed? Calculate the bandwith needed for the incoming bytes too. If you cannot feed the bytes fast enough to a tapedrive, the drive needs to stop/rewind/restart frequently, loosing lots of capacity, time, and will need replacement very soon. About the stated capacity: could it be you're somewhere in the terabytes of free space, and you're overflowing a 32bit long (size in KB used inside amanda)? Showing the exact output of the commands still helps a lot. -- Paul Bijnens, XplanationTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, ^^, * * F6, quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... * * ... "Are you sure?" ... YES ... Phew ... I'm out * ***
Re: Holding disk size misread by amcheck
LaValley, Brian E wrote: I have a Fedora Core 3 amanda server. I have specified an nfs mounted directory for one of the holding disks. Does anyone know why amcheck finds much less space available on this drive than a command like 'df' does? You need to explain a little more. "amcheck" checks only for free space (at least 64K) n AMANDA_TMPDIR, AMANDA_DBGDIR, and /etc. "amcheck" does not check space on the disklist entries. Show the messages that make you that. -- Paul Bijnens, XplanationTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, ^^, * * F6, quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... * * ... "Are you sure?" ... YES ... Phew ... I'm out * ***
Holding disk size misread by amcheck
I have a Fedora Core 3 amanda server. I have specified an nfs mounted directory for one of the holding disks. Does anyone know why amcheck finds much less space available on this drive than a command like 'df' does?