Hi, One possibility to skip the automatic tape medium detection is to add the following line at the beginning of /etc/udev.d/rules.d/60-persistent-storage.rules:
KERNEL=="btibm*", GOTO="persistent_storage_end" Michael Am Samstag, den 28.06.2008, 22:08 +0300 schrieb Juha Vuori: > Hi, > > there is still a kind of a problem in using 3592 tapes in SLES10-SP1, > seemingly related to the logical block device. > > When we do backups, we insert a tape media and then essentially run in a > script: > > vmcp att 500 '*' 181 > chccwdev -e 181 > > After that, however, all mt/mtst commands fail for a varying period of > time, depending on how much data there is on the tape. > > It seems that after setting the tape device online, first the tape > drived detects the amount of blocks on the tape: > > Jun 28 10:17:16 zlnx002 kernel: TAPE_BLOCK: Detecting media size... > ... > Jun 28 10:19:02 zlnx002 kernel: TAPE_BLOCK: Found 1559081 blocks on media > > during which the tape device is on IN_USE state operating FSF. > > After that the tape device operates for a while in BLKUSE state > operating BLK (I mean "lstape" tells that Op=BLK). Below is more > detailed log. > > I suppose the tape device initializes the first file on the disk to be > used as a block device automatically, which causes this lenghty delay in > other operations. > If I am correct, I think I would like to get the tape online only as a > locigal character device. Is there a way to do that? > We run udevd which makes all the tape device nodes available > automatically. Could this have something to do with the behaviour of the > block device? (I did not have time to test this, yet.) > > We can backup, though, using enough wait time before tarring or checking > lstape output in a loop, but this is ugly and I'd really like to > understand devices and drivers well... :) > > TIA, > Juha Vuori > > Our script: > ---------- > > # setupping first > # constants > TAPEDEV=/dev/ntibm0 > > # attach given tape device > /sbin/vmcp attach $1 '*' 181 2>&1 | logger -p local0.info > if [ $? -ne 0 ]; then > /sbin/vmcp msg op adminmsg - ATTACH $1 FAILED > exit 1 > fi > > sleep 1 > > # put the tape device online > /sbin/chccwdev -e 181 2>&1 | logger -p local0.info > if [ $? -ne 0 ]; then > /sbin/vmcp msg op adminmsg - $1 ONLINE FAILED > exit 1 > fi > > lstape | logger -p local0.info > > sleep 200 > > # set compression on > mtst -f $TAPEDEV compression 1 2>&1 | logger -p local0.info > > sleep 1 > > lstape 2>&1 | logger -p local0.info > > # rewind the tape > mt -f $TAPEDEV rewind 2>&1 | logger -p local0.info > > sleep 3 > > # put the data on tape using blocking factor 64 (means 32K blocksize) > tar --totals -b 64 -cf $TAPEDEV /bups/ 2>&1 | logger -p local0.info > > fails and produces log: > ---------------------- > > Jun 28 10:17:14 zlnx002 logger: TAPE 0502 ATTACHED TO ZLNX002 0181 > Jun 28 10:17:14 zlnx002 kernel: crw_info : CRW reports slct=0, oflw=0, > chn=0, rsc=3, anc=1, erc=4, rsid=27 > Jun 28 10:17:14 zlnx002 kernel: TAPE_CORE: tape device 0.0.0181 found > Jun 28 10:17:15 zlnx002 logger: Setting device 0.0.0181 online > Jun 28 10:17:16 zlnx002 kernel: TAPE_CORE: (0.0.0181): Tape has been mounted > Jun 28 10:17:16 zlnx002 kernel: TAPE_BLOCK: Detecting media size... > Jun 28 10:17:16 zlnx002 logger: Done > Jun 28 10:17:16 zlnx002 logger: TapeNo BusID CuType/Model > DevType/Model BlkSize State Op MedState > Jun 28 10:17:16 zlnx002 logger: 0 0.0.0181 3590/70 3590/10 > auto IN_USE FSF LOADED > Jun 28 10:19:02 zlnx002 kernel: TAPE_BLOCK: Found 1559081 blocks on media > Jun 28 10:20:36 zlnx002 logger: /dev/ntibm0: Device or resource busy > Jun 28 10:20:37 zlnx002 logger: TapeNo BusID CuType/Model > DevType/Model BlkSize State Op MedState > Jun 28 10:20:37 zlnx002 logger: 0 0.0.0181 3590/70 3590/10 > auto BLKUSE BLK LOADED > Jun 28 10:20:37 zlnx002 logger: mt: /dev/ntibm0: Device or resource busy > Jun 28 10:20:40 zlnx002 logger: tar: /dev/ntibm0: Cannot open: Device or > resource busy > Jun 28 10:20:40 zlnx002 logger: tar: Error is not recoverable: exiting now > > ---------------------------------------------------------------------- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
